Monazite and zircon U–(Th–)Pb dating reveals multiple episodes of HT metamorphism in the Cima Lunga unit (Central Alps): implications for the exhumation of high‐pressure rocks

Abstract

High- to ultrahigh-pressure (HP–UHP) rocks recording high-temperature (HT) > 700 °C are well exposed in the Central Alps, making it an ideal region to study the timing of metamorphic stages and the mechanisms of deep-seated rocks exhumation. Here, we report an integrated dataset of petrological and U–(Th–)Pb dating of metapelites surrounding ultramafic lenses from the Cima Lunga unit. At the interface with ultramafics preserving (U)HP–HT assemblages (1.5–3.1 GPa, 650–850 °C), metapelites record higher P‒T values (1.3–2.7 GPa, 700–850 °C) and traces of partial melting, whereas the rest of the unit is dominated by amphibolite-facies conditions. U–Th–Pb dating on zircon and monazite from migmatites indicates that partial melting was episodic involving at least two stages at ~38 to 35 Ma and 33–30 Ma, respectively. While the 38–35 Ma stage matches the HP conditions (> 1.5 GPa) and it is recorded around only one lens with scarce volumes of melt, partial melting at 33–30 Ma is witnessed at lower pressure (~1 GPa) and more widely distributed around the lenses, as within the major shear zones. Far from the ultramafics, zircon from the amphibolite-facies metasedimentary rocks record inherited pre-Variscan ages, while monazite ages at ~22 Ma document mineral growth during the Barrovian cooling. Field and petro-chronological evidence highlight that multiple episodes of partial melting locally developed at the rheological interface promoted by the interplay of fluids extracted from the ultramafic lenses associated with shear heating. New evidence suggests that local variation of P‒T equilibria play a significant role during the exhumation history.

Graphical Abstract

Introduction

Tectonic units with distinctive metamorphic assemblages can be used to reconstruct the thermal and mechanical evolution of orogens through Pressure–Temperature–Deformation–time (P‒T–D–t) paths (e.g., Grujic et al. 2002; Bousquet et al. 2008; Miyashiro 2012; Smye and England 2022; Vaughan-Hammon et al. 2022). However, structurally coherent tectonic nappes may preserve heterogeneous metamorphic records, locally marked by relatively small domains with relict (ultra)high-pressure ((U)HP) and/or high-temperature (HT) conditions embedded within lower-grade rocks (e.g., Evans and Trommsdorff 1978; Kaneko et al. 2000; O’Brien 2000; Liou et al. 2004; Zheng and Hermann 2014; Groppo et al. 2019). These sharp lateral variations in the P‒T record are key for investigating the complexities of the metamorphic history and the mechanisms driving the subduction–exhumation cycle (e.g., Burov et al. 2001; Hacker and Gerya 2013; Schmalholz et al. 2014; Agard 2021; Vaughan-Hammon et al. 2022). A major issue related to the interpretation of such contrasting metamorphic records is the identification of their origin due to tectonic complexity or consequence of geochemical processes. On the one hand, spatially variable metamorphic assemblages might result from different causes, such as chemical reactivity of protoliths, kinetics, spatial variability of fluids, differential retrogression of rocks sharing a common metamorphic history, or a combination thereof (Heinrich 1982; Wain 1997; Hacker et al. 2004; Zhou et al. 2022). On the other hand, heterogeneous metamorphic assemblages can also result from mechanical causes, such as (i) late to post metamorphic tectonic reworking and mixing (i.e., tectonic mélange; Hacker and Gerya 2013; Federico et al. 2007; Agard 2021) or (ii) tectonic control, local pressure and temperature variations decoupled from depth conditions (Schmalholz et al. 2014; Gerya 2015; Hess et al. 2022). To determine the geodynamics of these metamorphic records, it is critical to constrain the ages of the P‒T stages in rocks with different compositions. However, as heating close to the solidus involves larger entropy changes relative to pressure variation, the mineral assemblages easily lose their HP memory during the retrograde path (Proyer 2003). In contrast, the record of HT conditions (> 700 °C) is commonly much more persistent. Moreover, HT peak conditions can be captured by different geochronometers, such as zircon and monazite. Isotopic resetting becomes more striking as the fluid/melt availability increases, enhancing dissolution and recrystallization (Kohn 2016). However, temperature and pressure peaks are often diachronic, and if rocks exhume isothermally or with a T increase, the geochronometers likely experience retrograde conditions (e.g., Burton and O’Nions 1992). Zircon and monazite chemistry can be linked to distinct metamorphic assemblages and, thus, P–T stages, allowing us to constrain the metamorphic P‒T‒t path (Rubatto and Hermann 2007; Mottram et al. 2014; Kohn 2016; Kohn and Kelly 2018). Chemical exchange with the metamorphic assemblage, the effects of crystal–plastic deformation and interactions with fluids/melts exert fundamental control on the growth/resorption of these minerals, as well as on their internal elemental distribution (Rubatto 2002, 2017; Kohn et al. 2015; Kohn 2016; Piazolo et al. 2016; Corvò et al. 2023; Yakymchuk 2023). Correlating the isotopic ages with the P‒T stages requires a careful evaluation of the geochemical, textural, and deformation relationships between the geochronometers and the main mineral assemblage, as well as the fluids promoting both intracrystalline and intergranular element diffusion (Rubatto 2002, 2017; Kohn et al. 2015; Kohn 2016; Langone et al. 2018; Corvò et al. 2023).

In the European Alps, the (U)HP-nappes of the western sector of the orogen (e.g., Lago di Cignana, Zermatt-Sass, Monviso, Gran Paradiso and Monte Rosa) are characterized by relatively cold conditions (< 700 °C) at the pressure peak, followed by cooling during exhumation (e.g., Reinecke 1998; Agard et al. 2001; Angiboust et al. 2009; Groppo et al. 2009; Manzotti et al. 2015, 2018; Locatelli et al. 2018; Luoni et al. 2021). Even the archetypal UHP Dora-Maira massif experienced a similar evolution, although at slightly higher temperatures (700–750 °C; Compagnoni et al. 1995; Rubatto and Hermann 2001; Groppo et al. 2019; Manzotti et al. 2022). In contrast, the (U)HP rocks of the Central and Eastern Alps (Alpe Arami, Capoli, Monte Duria, southern Adula nappe, Cima Lunga unit, Pohorje unit) mostly show HT conditions (750–850 °C) at the metamorphic peak and thereafter for most of the retrogressive path up to ~1 GPa (Nimis and Trommsdorff 2001; Dale and Holland 2003; Brouwer et al. 2005; Janák et al. 2015; Tumiati et al. 2018; Pellegrino et al. 2020).

In this paper, we apply petrological and geochronological investigations to the Cima di Gagnone area (Cima Lunga unit, Central Alps, Switzerland; Fig. 1a, b), where the occurrence of ultra-high pressure and high-temperature ultramafic and lower-grade metasedimentary rocks has made this site a paradigmatic example of heterogeneous metamorphic record (Evans et al 1979; Pfiffner and Trommsdorff 1998; Corvò et al 2021; Piccoli et al 2022).

Fig. 1

Tectonic overview of the Lepontine area. a Location of the Lepontine dome in the Central European Alps context (modified after Burg and Gerya 2005). b Tectonic map of the eastern side (Ticino culmination) of the Lepontine dome (modified after Wenk 1955; Todd and Engi 1997; Brouwer et al. 2005; Steck et al. 2013; Cavargna-Sani et al. 2014; Corvò et al. 2021; Maino et al. 2021; Tagliaferri et al. 2023). c Geological map and cross section (from Tagliaferri et al. 2023). The Lepontine dome is rooted in nappes derived from the European distal margin. The Cima Lunga unit consists of a thin layer of metasedimentary rocks embedding orthogneisses, metacarbonate, and lenses of mafic and ultramafic rocks. Syn-tectonic metatexites and diatexites aligned along the upper thrust contact with the Maggia nappe are dated at ~32 Ma (Tagliaferri et al. 2023). Post-schistosity dikes and metamorphism associated with the migmatites parallel to the southern Alpine backstop are instead constrained at 22–24 Ma (Boston et al. 2017; Tagliaferri et al. 2023). The Cima di Gagnone area encompassing the (U)HP and HT occurrences is indicated with red boxes

Here, a complex pattern of heterogeneous P‒T record, varying from HP–HT to locally supra-solidus amphibolite-facies conditions, correlates with changes in composition and the structural position. However, the lack of robust chronological constraints prevents an unambiguous tectonic interpretation of the metamorphic stages. Determining the age and duration of the various P‒T stages is crucial for elucidating which tectonic model best fits the complex metamorphic pattern. Thus, we apply zircon and monazite U–Th–Pb geochronology to the metapelites of Cima di Gagnone, focusing on the migmatites locally occurring around the ultramafic lenses. We determine the trace element characteristics of the garnet, zircon, and monazite domains to constrain the conditions under which the accessory minerals formed and recrystallized. This allows us to correlate the various metamorphic assemblages with ages. The new data indicate that different metamorphic assemblages and textures correspond to different ages. Two distinct episodes of at T > 700 °C between ~38 and 35 Ma and ~33 and 30 Ma are selectively recorded by a few samples displaying different degrees of partial melting, whereas the volumetrically dominant, melt-absent, metapelites register only the late Barrovian cooling stage at ~22 Ma. The potential of fluids fluxing from the ultramafic lenses in promoting metasomatism and localized melting in the surrounding metapelites is addressed through thermodynamic modeling. Finally, the new data allow us to derive a critical review of the tectonic interpretations of the exhumation history applied to the Cima Lunga unit, recasting the results in the general Alpine framework.

Geological setting

The Central Alps started to be shaped in the Cretaceous, driven by the convergence and collision of the paleo-European and Adria tectonic plates, alongside the closure of the interposed Mesozoic Ligure-Piemontese and Valais basins (e.g., Stampfli et al 1998; Rosenbaum and Lister 2005; Handy et al. 2010; Zanchetta et al. 2012). In the Penninic domain of the Central Alps, which encompassing the paleo-European passive margin, the oceanic basins and the Briançonnais Peninsula—the transition from oceanic subduction to continent–continent collision occurred between the Eocene and the early Oligocene (e.g., Gebauer 1996; Schmid et al 1996; Wiederkehr et al., 2008, 2009). The following basement nappe stacking and the related metamorphic evolution are also well recorded in the Lepontine dome, which is a prominent structural and metamorphic dome composed of basement nappes derived from the European margin and the Valais basin (Froitzheim and Manatschal 1996; Schmid et al. 1996; Stampfli et al 1998; Maxelon and Mancktelow 2005). In the core of the dome, the Leventina and Simano nappes are flanked to the east by the Adula nappe and to the west by the Cima Lunga unit (Fig. 1b). These last two units preserve the best record of the (ultra)high-pressure conditions that appear as relicts in the dominant Oligocene amphibolite- to greenschists-facies Barrovian metamorphism delineated by the concentric distribution of its isograds cutting the nappe contacts (Fig. 1a, b; Wenk 1955; Merle et al. 1989; Steck et al. 2013; Tagliaferri et al. 2023).

The Cima Lunga unit is relatively thin (50–700 m thick from north to south), sandwiched between the overlying Maggia nappe and the underlying Simano nappe (Fig. 1c; Tagliaferri et al. 2023). The unit consists of pre-Variscan metasedimentary rocks intruded by Ordovician–Permian granitoids (Tagliaferri et al. 2023). This sequence envelops lenses (10 cm–1 km in length) of peridotites and eclogites/amphibolites. (Ultra)mafic rocks record Alpine (U)HP-HT metamorphism (up to 3.0 GPa and 800 °C), whereas metapelites and orthogneisses generally exhibit amphibolite-facies conditions (⁓0.8 GPa, 650 °C) (Fig. 1b; Grond et al. 1995; Pfiffner and Trommsdorff 1998; Pfiffner 1999; Nimis and Trommsdorff 2001; Brouwer et al. 2005; Scambelluri et al. 2014; Cannaò et al. 2015). However, minor (U)HP–HT relicts (1.5–2.7 GPa, 750–800 °C) have also been recently found in metapelites (Corvò et al. 2021; Piccoli et al. 2022). These relicts occur in narrow (cm to m thick) and discontinuous rims of migmatites developed around peridotite lenses (Fig. 2a, b). Geochemical signatures of the migmatitic metapelites attest for a strong interaction with fluids sourced from the adjacent mafic and ultramafic rocks (Pfeifer 1987; Heinrich 1982; Früh-Green 1987; Pfiffner 1999; Corvò et al. 2021). Similar migmatites can be discontinuously found also along the main thrust dividing the Cima Lunga from the overlying Maggia nappe (Fig. 1c; Tagliaferri et al. 2023). The volumetrically dominant paragneisses and orthogneisses of the rest of the northern Cima Lunga unit apparently lack any evidence of partial melting. Structural investigations have described long-lasting coupled deformation of peridotites and their metapelitic matrix under high bulk shear strain with γ > 14.5 (Maino et al. 2021). A large rheological contrast led to the development of heterogeneous deformation patterns, including sheath folding in the weak metapelitic matrix, while the more rigid (ultra)mafic lenses experienced fracturing and rotation of the internal layering, as revealed by multiple folding and boudinage stages (Maino et al. 2021).

Fig. 2

a, b Sample location of the metasedimentary rocks collected around two ultramafic lenses in Cima di Gagnone (Cima Lunga unit): a coincide with Mg31 outcrop and b is Mg160 outcrop, respectively, after Pfiffner and Trommsdorff (1998) and Corvò et al. (2021). c Panoramic view of (ultra)mafic (UM), metasedimentary rocks and evidence of large migmatite patches at Passo di Gagnone (between Mg31 and Mg32 after Pfiffner and Trommsdorff 1998). Migmatites develop at the interface between metapelites and mafic/ultramafic lenses or as infill of large fracture within the (ultra)mafic lenses. d Detail of the ultramafic–migmatite contact: migmatitic gneiss contain abundant fragment of mafic relicts showing cuspate–lobate textures (arrows) and are bordered by thick leucocratic rims. e Detail of the progressive transition between migmatitic gneisses (encompassing blocks of mafic rocks) and metapelites, the latter still containing significant amount of leucocratic segregated melt (arrows). f Detail of stromatic migmatites, partially involved within the schistosity (Mg31; Passo di Gagnone). g Folded and sheared migmatitic metapelites with leucocratic intrafolial pods (Mg31; Passo di Gagnone)

Geochronological background

Zircon U–Pb sensitive high-resolution ion microprobe (SHRIMP) ages between 43 ± 2 Ma and 35.4 ± 0.5 Ma and garnet Sm–Nd ages of 40 ± 4 Ma are interpreted as constraining the timing between the (U)HP stage and the initial decompression, still at HP conditions, of peridotites and eclogites of the Cima di Gagnone area (Becker 1993; Gebauer 1994, 1996, 1999). Neighboring metasedimentary rocks mostly record Paleozoic U–Pb zircon ages (Gebauer 1994, 1996, 1999; Corvò et al. 2021; Tagliaferri et al. 2023), although a few zircon rims from both peridotite and metasedimentary rocks register younger ages at 33–30 Ma (Gebauer 1994, 1996; Berger et al. 2005; Corvò et al. 2021; Tagliaferri et al. 2023). This time span is well recorded by syn-kinematic migmatites, which occur scattered within the shear zone between the Cima Lunga and Maggia nappe (Fig. 1c; Tagliaferri et al. 2023). Noticeably, two zircon grains from the metapelites preserving HP-HT relicts and evidence of partial melting revealed U–Pb zircon ages between 37 ± 4 Ma and 34 ± 5 Ma (with a weighted average age of 36 ± 1.2 Ma) from cores to rims (Corvò et al. 2021). However, these data are poorly correlated with the chemistry of the geochronometer and the host mineral assemblage, thus providing only a broad constraint on the age of peak metamorphism.

Geochronological data from the Cima Lunga unit are consistent with the ages of the other U)HP-HT occurrences in the Central Alps (Adula nappe, Alpe Arami and Monte Duria) where eclogites and garnet peridotites yield U–Pb, Lu–Hf and Sm–Nd ages between 45 and 34 Ma, which are interpreted to date the Alpine (U)HP stage and subsequent decompression still at HP conditions (Gebauer 1996; Brouwer et al. 2005; Hermann et al. 2006; Herwartz et al. 2011). U–Pb SHRIMP and laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) zircon data from eclogites and their metasedimentary country rocks from central and northern sectors of the Adula nappe instead yield dominant Paleozoic ages (330–340 and 370 Ma), with a few ages of approximately 33–32 Ma from zircon rims interpreted as being associated with the regional nappe emplacement (Liati et al. 2009; Sandmann et al. 2014; Tagliaferri et al. 2023).

Zircon, monazite, and allanite ages of ~32 and ~22 Ma, respectively, from the southern Lepontine dome are interpreted to date a long-lasting phase of Barrovian thermal re-equilibration (~10 Ma, Köppel and Grünenfelder 1979; Rubatto et al. 2009; Tagliaferri et al. 2023) encompassing multiple magmatic/migmatitic events along the Insubric Line (Fig. 1b, c; Gregory et al. 2012; Tagliaferri et al. 2023). At the same at the northern margin of the dome, the same geochronometers suggest the onset of renewed nappe stacking (Janots et al. 2009; Gregory et al. 2012; Boston et al. 2017).

The debated tectonic evolution of the Cima Lunga unit

The contrasting metamorphic record between the (U)HP lenses of peridotites and eclogites and the surrounding orthogneisses and metasedimentary rocks has been interpreted for a long time as being due to tectonic mixing of exotic blocks with gneisses and schists within a subduction channel mélange (Trommsdorff 1990; Engi et al. 2001; Berger et al. 2005; Scambelluri et al. 2015). This model has recently been challenged by (i) the recognition of the structural coherence of the unit (Maino et al. 2021; Tagliaferri et al. 2023), (ii) the determination of a pre-Variscan age of the metasedimentary sequence (Tagliaferri et al. 2023), and (iii) the finding of continuous pressure and temperature gradients in metapelites, reaching HP and HT conditions similar to those recorded by the Cima Lunga unit (up to 2.7 GPa and 850 °C) in narrow halos that discontinuously surround the ultramafic lenses (Corvò et al. 2021; Piccoli et al. 2022). These HT–HP conditions are apparently localized along the lithological interfaces or within high-strain zones such as the tectonic boundary at the top of the nappe, and often accompanied by intense melt-rock interaction (Corvò et al. 2021; Tagliaferri et al. 2023). Evidence of partial melting is instead absent, or very limited, in the rest of the unit, thus it is debated if these HT conditions reflect peak pressure conditions or decompression. More importantly, it is also debated if the HT conditions were experienced by the entire unit but are preserved only locally, or, alternatively, if they developed only at the metapelite–ultramafic rocks interface, as well as along the major shear zones. Contrasting interpretations result in different exhumation models, invoking either coexisting heterogeneous metamorphic equilibria along rheological interfaces (Corvò et al. 2021) or differential preservation of the P‒T record during the retrogressive path (e.g., Piccoli et al. 2022).

Field relationships and sample description

Geochronological studies were performed on five selected zircon/monazite-bearing samples for which Corvò et al. (2021) described the petrography, microstructure and P‒T evolution. All the samples are garnet-bearing metapelites characterized by alternating mica- and quartz/plagioclase-rich layers.

Sample M92 occurs at a distance of ⁓50 m from the ultramafic lens (Fig. 2a, outcrop Mg31 of Pfiffner and Trommsdorff 1998) and does not show any evidence of partial melting (Corvò et al. 2021). For this reason, sample M92 is taken as representative of melt-free metapelites representing the dominant part of the Cima Lunga unit. These rocks show a homogeneous assemblage of plagioclase + quartz + biotite + muscovite + garnet + kyanite ± staurolite (see Figs. 3H, 4A from Corvò et al. 2021). Sample M92 preserves two overgrowths of almandine-rich garnet both containing inclusions of quartz, plagioclase, muscovite, biotite and rutile (see Figs. 5A–D and Fig. 6 from Corvò et al. 2021). Cores are significantly richer in inclusions than the overgrowths. Larger garnets contain mainly quartz inclusions and parts of the garnet have been replaced by quartz, suggesting that the garnet has been partly resorbed. Thermobarometry and thermodynamic modeling point to amphibolite-facies conditions for both garnet cores and overgrowths (P < 1.1 GPa and T < 700 °C; Heinrich 1982; Grond et al. 1995; Pfiffner 1999; Corvò et al. 2021). As the inclusion trails trapped within the garnet cores are folded and truncated by the overgrowth, the cores are considered pre-kinematic, likely formed during the Variscan metamorphism that affected the paragnesses precursors of the Lepontine basement nappe (Rutti 2003; Rutti et al. 2005; Liati et al. 2009; Cavargna-Sani et al. 2014; Tagliaferri et al. 2023). On the other hand, garnet overgrowths appear to be syn-kinematic with the main foliation and in microstructural equilibrium with the matrix assemblage; thus, they are considered Alpine in age.

Fig. 3

Back-scattered images and X-ray elemental maps (Y and Th; hc = high concentration; lc = low concentration) of representative monazite grains; a–c refer to monazite from Group I (samples M92, MP3, M102), while d–h to Group II (samples M69 and M119; see Table S5 in the Suppl. Mat.). Yellow and blue circles represent the LA-ICP-MS spot analyses (10 µm diameter) and trace element analyses (25 µm diameter), respectively. The relative U–Pb (italic), U–Th–Pb (bold italic) and ²⁰⁶Pb/²³⁸U (normal) dates are shown (Ma) with 2σ error. Mineral abbreviations are after Whitney and Evans (2010)

Fig. 4

Monazite U–Th–Pb results. a Simplified sketch showing the structural position of analyzed samples with respect to the neighboring ultramafic lenses (Mg31 and Mg160 as in Fig. 2a–b; after Corvò et al. 2021). The interface between metapelite and ultramafic lenses (halos) displays evidence of fluid-mediated chemical exchange (dark brown) and partial melting (gray). For each sample, metamorphic estimates (from Corvò et al. 2021) and Y-content of monazite grains are shown. b Uncorrected data plotted in Tera-Wasserburg concordia space for the two groups of monazite (Group I: M92, MP3, M102; Group II: M69, M119). Model-I linear regression (black lines) was conducted in IsoplotR (Vermeesch 2018) to obtain the lower intercept. The dashed and solid lines indicate the free-regression and the anchored-regression (Stacey and Kramer 1975), respectively. Lower-intercept age is from the anchored-regression (SK-anch.). Empty ellipses are not considered in the calculation. c ²⁰⁶Pb/²³⁸U vs. ²⁰⁸Pb/²³²Th diagrams (Vermeesch 2020) for samples showing discordance between U–Pb and Th–Pb ages < 4%. Dashed empty ellipses indicate the discordant data excluded for the calculation of the U–Th–Pb Concordia age. The ages are shown with 2σ error

Fig. 5

Cathodoluminescence and backscattered images of representative zircon grains from samples M92 (a–c), M102 (d–g) and M119 (h–o). Yellow circles represent the LA-ICP-MS spot analyses (10–25 µm diameter) with the relevant concordant U–Pb ages (Ma with 2σ error). The dashed yellow circles show the LA-ICP-MS spot analyses that gave discordant U–Pb data

Fig. 6

Zircon U–Pb results. a Simplified sketch showing the structural position of analyzed samples with respect the neighboring ultramafic lenses as in Fig. 4a. b Plot of the concordant U–Pb data and the relative Weighted Average (WA) age for well-defined zircon populations. White-filled bars refer to U–Pb zircon outliers mostly associated with cores or altered rims of large grains, excluded for the weighted average age calculation (samples M102, MP3 and M92); unfilled bars refer to zircon included in garnet (sample M119). N: number of analyses. c Wetherill plots of the U–Pb data represented as colored bars in b. Concordia ages from samples M119 and M102 are displayed. Dashed empty ellipses indicate data excluded from the calculation of the U–Pb Concordia age. Ages are shown with 2σ error

The other four samples (samples MP3, M102, M69 and M119) were collected around the ultramafic lenses, where metasedimentary rocks grade locally to variably thick layers of metatexites or sporadic diatexites (Fig. 2a–c). We stress that the ultramafic lenses are not completely enveloped by migmatites; rather, these lenses occur as a discontinuous and quite heterogeneous layer with variable thickness. Migmatites may be thin (cm-sized), exposed as large (m-sized) patches filling fractures and other dilatant sites or pressure shadows (Fig. 2c–e). Partial melted rocks typically contain leucosomes and melanosomes, which are either stromatic or aligned and shared along the main schistosity (2f–g). The migmatitic sequence also contains fragments of mafic or ultramafic rocks (Fig. 2d, e). These blocks show cuspate–lobate contacts with the embedding migmatitic gneisses. Moreover, the thick, white plagioclase-rich leucosomes often mark pressure shadows around the mafic blocks, highlighting the flow control on the melt distribution (Fig. 2d, e). Migmatites may be folded together with the melt-free metapelites testifying pre- to syn-deformation partial melting (Fig. 2f, g).

The sampled rocks display variable melt proportions, with the minimum volume of melt in proximity to the ultramafic lenses estimated to be between 0.5 and 8 vol.% (Corvò et al. 2021). Generally, these samples show small variations in their mineral assemblage (e.g., absence of staurolite) with respect to the metapelite far from the ultramafic lenses (Corvò et al. 2021). However, they show increasing abundances of feldspar in all migmatites and quartz becomes rare in sample M119 (Fig. 7A–D from Corvò et al. 2021). In all the samples, garnet crystals up to 2 mm in diameter contain inclusions of quartz, plagioclase, biotite and muscovite (or phengite). Compelling evidence of partial melting include quartz + plagioclase leucosomes and typical melt-rock microstructures (e.g., cusps, lobes and string of beads; Fig. 7E from Corvò et al. 2021). Sample M119 has a peculiar Al₂O₃-rich and SiO₂-poor composition and displays a coarse grain size, low crystallographic preferred orientation of minerals and weak intracrystalline garnet deformation.

Fig. 7

P–T pseudosections and isomode of expected melt (vol.%) for samples M119 (a, b), M92 (c, d), MP3 (d, e) and D2 (g, h). Phase diagrams are reproduced from those presented in Corvò et al. (2021) (M119, M92 and MP3) and Piccoli et al. (2022; Bulk D2). Filled red, orange, blue, and white stars indicate the peak conditions inferred for samples M119, M92, MP3 and D2. Empty dotted red, orange, blue and black stars represent the projection of the corresponding sample in the other phase diagrams. Green, red, pink, and pale blue circles indicate the prograde, peak, early, and late retrograde steps described for sample D2 (Piccoli et al. 2022)

The microstructures of the migmatitic samples are broadly consistent with high temperatures between 700 and 850 °C, being the highest recorded by sample M119 (Corvò et al. 2021). However, such an HT record correlates with variable P estimates from the different samples (from ⁓0.8 to 1.1 GPa of sample MP3 to 1.4–1.7 GPa of samples M102 and M119). However, higher-pressure conditions up to 2.5 GPa may eventually be inferred in samples M102 and M119 based on the Si content of phengite (Heinrich 1982; Cannaò et al. 2015; Corvò et al. 2021). Piccoli et al. (2022) described the P‒T evolution of a metapelite (samples D1 and D2) very similar to sample M119 in terms of composition, texture and petrography, which was also collected adjacent to the same garnet peridotite lens of Cima di Gagnone (outcrop Mg160 in Pfiffner and Trommsdorff 1998). The mineral record of these samples (D1 and D2) has been interpreted as describing an evolution from prograde to HP-HT peak conditions (2.7 GPa and 800 °C) followed by cooling still under HP conditions (~2.5 GPa and 700–750 °C) prior to decompression and further cooling at 1.0 GPa and ~620 °C (Piccoli et al. 2022).

Results

Monazite distribution and texture

We recognized two groups of monazite grains in different samples. Group I (samples M92, MP3 and M102) is mostly composed of tiny (< 20 μm), elongated monazite grains, which are too small to be analyzed. Some larger subhedral crystals (20–100 μm in length) occur in association with biotite–muscovite layers (Fig. 3a–c). Some of them display straight grain boundaries, while others have irregular boundaries and are partially replaced by allanite (Fig. 3c). Monazite grains from group II (samples M69 and M119) occur as relicts (250–500 μm in length) rimmed by allanite, which are in turn surrounded by clinozoisite (Fig. 3d–h). Group II monazite crystals exhibit irregular, inwards-penetrating and cuspate–lobate grain boundaries (Fig. 3d–h). Corona structures are locally associated with or included in garnet, but the original monazite-garnet relationships are obliterated by subsequent allanite replacement (Fig. 3e). High-contrast backscattered (BSE) imaging reveals that the monazite grains of both groups are nearly homogenous; somewhat brighter zoning is locally restricted to rims or fractured zones (Fig. 3).

Monazite chemistry and trace element compositions of monazite-garnet pairs

Microprobe analyses revealed significant chemical differences between the two monazite groups (Table S1). The grains of group I (samples M92, MP3 and M102) show moderate contents of Y₂O₃ (1.8 ± 0.2 wt.%), Sm₂O₃ (2.4 ± 0.2 wt.%), Gd₂O₃ (1.9 ± 0.2 wt.%), Dy₂O₃ (0.6 ± 0.2 wt.%) and ThO₂ (3.3 ± 1.2 wt.%) but are significantly greater than those of group II (samples M119 and M69), which are extremely poor in Y₂O₃ (0.03 ± 0.03 wt.%), Sm₂O₃ (1.7 ± 0.2 wt.%), Gd₂O₃ (0.8 ± 0.2 wt.%) and Dy₂O₃ (0.04 ± 0.05 wt.%). On the other hand, group II monazite is enriched in ThO₂ (5.4 ± 0.6 wt.%). Within group I, the Y content does not show an evident core-rim trend (Table S1). Th concentrations are generally higher corresponding to rims or fractures (Table S1). The U content is inversely correlated with Th, resulting in a core-to-rim increase in the Th/U ratio (from 6.6 ± 3.7 and 9.7 ± 5.7 for groups I and II, respectively).

Elemental X-ray mapping of Y and Th from group I monazites shows a patchy distribution (Fig. 3a–c). The group II monazite grains are instead nearly homogenous for all elements, except for Th, which significantly increases in the rims or along fractures and necked zones (Fig. 3d–h). Th zoning overlaps the bright domains revealed by BSE imaging.

To investigate the relationships between monazite (group II) and garnet in sample M119, we performed trace element analyses on neighboring monazite–garnet pairs (Tables S2–S3; Fig. S1). Monazite grains display high Sr concentrations (1310 ± 367 ppm) and weak negative Eu anomalies (Eu_N/Eu*_N ~0.13) and are depleted in heavy rare earth elements (HREEs) (Fig. S1; Table S2). The trace element contents of garnets are nearly homogeneous from core to rim (Fig. S1; Table S3). Garnet crystals generally have low-Y contents (190 ± 64 ppm) and flat middle-HREE (M-REE) trends (Dy_N/Yb_N = 1.15 ± 0.3; Dy_N/Gd_N = 1.57 ± 0.5). The partition coefficients for rare earth elements (REEs) between monazite and garnet rims or cores (REE_Dmnz/g) were calculated and represented in an array plot, as proposed by Taylor et al. (2017) and Manzotti et al. (2018) (Fig. S1). Monazite-garnet cores display a linear trend, whereas the partitioning of monazite with garnet rims and fragments appears to be more random.

Monazite U–Th–Pb geochronology

Five analyses of four monazite grains from sample M92 yield two concordant U–Pb ages (e.g., with U–Pb discordance <  ± 10%), at 121 ± 3.5 Ma and 22 ± 0.9 Ma, while only one of them resulted concordant also for the Th–Pb system (with discordance between U–Pb and Th–Pb data <  ± 4%), at 21.9 ± 0.3 Ma (Table S4; Fig. 3a). The complexities encountered for this sample, such as the rare occurrence of monazite, the small dimension of grains, and a poor concordance among the isotopic systems prevent the calculation of a robust geologically meaningful age. The uncorrected monazite data form a linear array in the Tera–Wasserburg diagram (Fig. 4b). Both free- and anchored regression of these data produced nearly identical intercept dates at 21.8 ± 0.5 Ma and 21.9 ± 0.5 Ma, respectively (MSWD = 1.6; Table S4; Fig. 4b).

Sixteen analyses on nine crystals of sample MP3 yield nine concordant U–Pb data, highlighting two populations at 430 ± 9–418 ± 8 Ma and 23 ± 1–20.7 ± 0.7 Ma (Table S4; Fig. 3b). Only one data point shows good U–Th–Pb concordance at 22.1 ± 0.6 Ma. In this sample, the pre-Alpine domains have slightly lower Y- and Th contents than the Alpine domains. In the Tera–Wasserburg space, the uncorrected monazite data form a discordant, linear array (Fig. 4b). The free- and anchored regressions produce identical intercept dates at 21.5 ± 0.3 Ma (MSWD = 2.2; Table S4; Fig. 4b).

Five analyses on five crystals of sample M102 yield only one U–Pb concordant age of 23 ± 0.7 Ma. In Tera–Wasserburg diagram, the uncorrected monazite data do not form a well-defined linear array (Fig. 4b). By excluding one ratio, the free-regression produced a slightly younger lower intercept date (21.4 ± 1 Ma; MSWD = 0.58) than the anchored-regression one at 22.3 ± 0.5 Ma; MSWD = 5; Fig. 4b). The discordance between the U–Pb and Th–Pb ages is instead <  ± 4% in four out of five analyses in the range of 21.5 ± 0.3–24.6 ± 0.6 Ma (Figs. 3c, 4c). In the ²⁰⁶Pb/²³⁸U versus ²⁰⁸Pb/²³²Th diagram, the data plot as cluster with a Concordia age of 21.9 ± 0.4 Ma (MSWD = 2.8; Fig. 4c).

The geochronological signal of the Y-depleted monazite of group II (Y < 0.1 wt.%) differs significantly. Eight U–Pb analyses on two crystals of sample M69 resulted discordant defining a linear array in the Tera–Wasserburg diagram (Table S4; Fig. 4b). The free regression of isotopic data provides a slightly older lower intercept date (33.6 ± 4.7 Ma, MSWD = 0.3) than the anchored-regression one at 33.0 ± 0.4 Ma (MSWD = 0.3; Fig. 4b. Although the high discordance between the two U–Pb systems, the concordance between ²⁰⁶Pb/²³⁸U and ²⁰⁸Pb/²³²Th systems is always >  ± 96% and provide a Concordia age of 37.1 ± 0.7 Ma (MSWD = 0.1; Figs. 3d, 4b).

Twenty analyses on six crystals of sample M119 yield discordant U–Pb results defining a linear array in Tera-Wasserburg space (Fig. 4b). The free-regression of those data produced a slightly younger lower intercept date (34.4 ± 0.7 Ma, MSWD = 1.6) than the anchored-regression date at 35.0 ± 0.3 Ma (MSWD = 1.6; Fig. 4b). Although the U–Pb systems provided only discordant data, the concordance between the ²⁰⁶Pb/²³⁸U and ²⁰⁸Pb/²³²Th data is higher (8 of 20). These U–Th–Pb concordant dates range from 40.8 ± 0.6 to 36.2 ± 0.6 with a Concordia age of 38.2 ± 0.7 Ma (MSWD = 0.2; Fig. 3e–h, 4b). Noticeably, the Th-rich rims and near-fracture zones of these monazite grains correspond to the youngest dates < 38 Ma, whereas the innermost monazite domains preserve the older dates > 38 Ma (Fig. 3e–f).

In summary, monazite crystals with moderate-Y contents (group I, Y values from 1.3 to 2.3 wt.%) provide evidence for an Alpine event at ~21 to 22 Ma. Differently, the Y-poor monazite grains of group II (Y < 0.1 wt.%), record only an older Alpine population with U–Pb lower intercepts dates between 33 and 35 Ma and ²⁰⁶Pb/²³⁸U versus ²⁰⁸Pb/²³²Th concordia ages of ~37 to 38 Ma. Moreover, pre-Alpine dates are only preserved in the monazite grains from rocks collected at the boundary or outside the interface with the ultramafic lenses (Fig. 4a).

Zircon distribution and texture

Zircon grains of samples M92 and M102 are mainly prismatic in shape, with variable lengths up to 250 μm (Fig. 5a–g). Cathodoluminescence (CL) images reveal inherited cores with sector or oscillatory zoning wrapped by rims with variable features. In sample M92, rims are absent or very thin (< 10 μm) with bright to gray CL emission (Fig. 5a–c). In contrast, in sample M102, the rims are 10–30 μm thick, gray, and have low luminescence and faint zoning (Fig. 5d–g). Sample MP3 mainly contains zircon grains with bright and thin rims, although a few grains with gray or oscillatory-zoned thick rims occur (Corvò et al. 2021).

Sample M119 contains zircons both in the matrix and as inclusions in garnet (Fig. 5h–o). Two large prismatic grains (up to 400 μm long) within the matrix exhibit cores with broad banded zoning characterized by variable CL brightness, from black to bright. Relatively thick gray rims (10–30 μm; Fig. 5h–i, m–n) surround the zircon cores. One smaller rounded crystal shows a zoned core discordantly mantled by a low-luminescence thick (30 μm) rim (Fig. 5o). The small zircon inclusions within garnet (up to 50 μm in length) are characterized by bright cores surrounded by darker rims (Fig. 5j, k). One basal section of a zircon grain partially included within the rim of a garnet (Fig. 5j, l) displays a homogeneous gray texture characterized by a fracture mimicking the garnet edge.

Zircon trace element compositions and Ti-in-zircon thermometry

REE patterns from the oscillatory zoning domains of zircons from sample M92 (Table S5; Fig. S2a) show a pronounced negative Eu anomaly, relatively steep pattern of M-HREEs (Lu/Gd_N from 49 to 142) and Th/U of 0.25 ± 0.09, suggesting magmatic growth (Hoskin and Schaltegger 2003; Rubatto 2017). The Ti content of zircon is 7.3 ± 1.4 ppm. Using the Ti-in-zircon thermometer of Ferry and Watson (2007) and assuming TiO₂ and SiO₂ saturation, this temperature is 716 ± 18 °C (Table S5). However, the reduced thickness of the rims (< 10 μm) hampers the distinction between rims and cores.

REE patterns from the oscillatory-zoned domains of zircons from sample M102 show a strong Eu anomaly (Eu/Eu* = 0.09 ± 0.01), moderately steep M-HREE pattern (Lu/Gd_N between 16 ± 2) and Th/U of 0.29 ± 0.16, whereas the rims are approximately one or two orders of magnitude poorer in REEs, lack Eu anomalies (Eu/Eu* = 0.67 ± 0.18) and have Th/U ratios < 0.1 (Table S5, Fig. S2a). These features indicate the growth of metamorphic rims on inherited magmatic domains (Hoskin and Schaltegger 2003; Rubatto 2017). The Ti content is 7.4 ± 1.2 ppm without a significant difference between rims and magmatic overgrowths, indicating a temperature of 718 ± 15 °C (Table S5).

Two large zircon grains from sample M119 were characterized for trace element concentrations (Fig S2b, Table S5). Cores and rims are depleted in Pb and Th, with Th/U < 0.1. Both intragrain domains show heterogeneous contents of Y, U, and REEs and HREE slopes (Lu/Gd_N values from 4 to 62), without systematic correlation with textural position. The Eu anomaly generally increases from core to rim (Eu/Eu* = 0.46 ± 0.3 and 0.25 ± 0.16, respectively; Fig S2b, Table S5).

The Ti contents ranges from 7.9 to 12.8 ppm in the cores, and from 6.0 to 7.6 ppm in the rims. Ti-in-zircon thermometry indicates that the rims and cores of the zircon from sample M119 have temperatures of 712 ± 21 °C and 751 ± 18 °C for rims and cores, respectively (Table S5).

Zircon U–Pb geochronology

Thirty-two analyses of seventeen zircon grains from sample M92 yield twenty-four concordant U–Pb ages ranging between 579 ± 15 Ma and 162 ± 5 Ma (Figs. 5a–c, 6b; Table S6). Concordant ages from the oscillatory zoning are between 484 ± 13 Ma and 422 ± 9 Ma (Fig. 6b, c). Among the discordant ages, two thin bright rims have ²⁰⁶Pb/²³⁸U data of 32 ± 1 Ma and 33 ± 1 Ma.

A dominant pre-Alpine population is also recorded in sample MP3 (analyses reported in Corvò et al. 2021; Table S6), which provides concordant data for all the zoning patterns spanning between 533 ± 9 and 384 ± 7 Ma, with a narrow cluster for concordant data from the oscillatory zoning (533 ± 9 and 461 ± 8 Ma; Fig. 6b, c). Only one thick homogeneous rim yields a U–Pb age of 31.5 ± 0.6 Ma (Fig. 6b, c).

Twenty-four analyses of three zircons from sample M102 yield twenty-three concordant data points spanning between 686 ± 20 Ma and 30.2 ± 1.9 Ma (Figs. 5d–g, 6b; Table S6). The oscillatory zoning domains are mostly pre-Variscan with dates between 686 ± 20 Ma and 555 ± 18 Ma (Fig. 5e–g). Eighteen spots from the outer thick metamorphic rims truncating the magmatic overgrowths show data scattered between 360 ± 21 Ma and 30.2 ± 1.9 Ma (Fig. 5e–g). Among these, the youngest population of eight zircons spanning between 33.2 ± 2.3 Ma and 30.2 ± 1.9 Ma yields a weighted average U–Pb age of 31.4 ± 1.6 Ma (MSWD = 0.2; n = 8; Fig. 6b) and an identical Concordia age of 31.4 ± 0.8 (MSWD = 0.1; n = 8; Fig. 6c).

Twenty-five analyses of three zircon grains from the matrix of sample M119 yield twenty-four concordant U–Pb ages ranging between 37.5 ± 3.3 Ma and 32.7 ± 3.1 Ma (Table S6; Fig. 5h–o), with a weighted average U–Pb age of 34.9 ± 0.8 (MSWD = 0.9; n = 24; Fig. 6b) and an identical Concordia age of 34.9 ± 0.6 (MSWD = 1.3; n = 17; Figs. 5, 6c). A similar U–Pb Concordia age of 35.9 ± 0.6 Ma results from two other crystals of sample M119 analyzed in Corvò et al. (2021) (Table S6). Isotopic data show a good correlation with textural position and intragrain position of analytical spots. In fact, although mostly within the analytical error, data from the outer rims record somewhat younger dates (~32 to 35 Ma) with respect to the internal overgrowths (~35 to 37 Ma; Fig. 5i–o). Further six analyses have been performed on two zircon grains partially or totally included within garnet record older data. Two analyses on one crystal entirely included in garnet yield a discordant U–Pb core (with a ²⁰⁶Pb/²³⁸U data point of 134 ± 3 Ma) mantled by a relatively thick rim with a concordant U–Pb age of at 41.2 ± 0.8 Ma (Figs. 5j–k, 6b). Four analyses of a grain partially included in garnet supply four concordant U–Pb data between 44.8 ± 4.3 Ma and 37.1 ± 4.7 Ma (Figs. 5j, l, 6b).

Thermodynamic modeling of the expected melt

P‒T phase diagrams of representative samples (M119, M92 and MP3 from Corvò et al. 2021 and D2 from Piccoli et al. 2022) provide peak pressure and temperature conditions of 820 °C and 1.5 GPa for M119, 660 °C and 1.1 GPa for M92, and 700 °C and 1.0 GPa for MP3 (Fig. 7). In this study, the pseudosections are integrated with curves representing the modal amount of melt (vol.%) expected in the range of P‒T conditions of the model (Fig. 7). The peak conditions of sample M92 plot in the melt-out part of the diagram (Fig. 7c, d) and consistently show no evidence of melting, whereas the P‒T conditions of the migmatitic samples (MP3 and M119) match increasing melt volumes of ~4 to 10 vol.% (Fig. 7a, b, e, f). On the other hand, despite preserving evidence of melting, the P‒T conditions of the peak (800 °C, 2.6 GPa) calculated for sample D2 never reach the solidus (Fig. 7g, h; Piccoli et al. 2022). A variable amount of melt from approximately 10–30 vol.% is expected in all the samples under P‒T conditions recorded by the migmatitic samples (MP3 and M119, respectively, the blue and red dotted empty stars in Fig. 7a, b, e, f), assuming these samples have shared a common metamorphic history.

Discussion

Monazite age interpretation

Monazite is a very good tracer of petrological processes because it exhibits textures that directly reflect metamorphic reactions (Bosse and Villa 2019). The correlation between chemistry and age in monazite is well constrained, particularly when rocks experience partial melting (e.g., Pyle and Spear 1999, 2003; Pyle et al. 2001; Kohn et al. 2005; Kelsey 2008; Rubatto et al. 2013; Kohn 2016; Varga et al. 2020). In general, the Th and Y contents in monazite decrease during prograde metamorphic reactions in the presence of garnet, until partial melting causes monazite dissolution (Rapp et al. 1987; Varga et al. 2020). After melt crystallization, new high-Y and high-Th overgrowths on relict grains are usually expected (Pyle and Spear 2003; Rubatto et al. 2013; Casini et al. 2023). As a rule of thumb, monazite chemical domains correlate with age population following this scheme (Kohn 2016): early prograde (< 700 °C, high-Y and high-Th cores), late prograde/early pre-melting (> 700 °C, low-Y and low-Th growths) and post-melting (< 700 °C, high-Y and high-Th rims). At lower temperatures (< 600 °C), monazite can be completely or partially replaced by allanite, clinozoisite and apatite (Spear 2010) or, modified by fluids. Fluid-mediated coupled dissolution–precipitation processes may cause further redistribution of elements, typically resulting in U depletion and Th enrichment along reactive interfaces such as grain boundaries, rims, or fractured zones (e.g., Bosse et al. 2009; Seydoux-Guillaume et al. 2012). The preservation of prograde ages in monazite is also a function of crystal size, as the largest monazite grains are less consumed during melting and subsequent retrogression (Langone et al. 2011).

Data from the Cima di Gagnone metapelites, show an evident discordance in the ²⁰⁷Pb/²³⁵U versus ²⁰⁶Pb/²³⁸U ratios in the Tera-Wasserburg concordia space for all monazite domains. This discordance may derive from several possible causes: mixing of domains, intermediate daughter product disequilibrium, the presence of initial Pb and the mobilization of Pb. The provided careful textural and chemical characterization and mapping of the monazite domains limits the first cause as a possible explanation. Instead, the incorporation of ²³⁰Th (Parrish 1990), disturbance of the original initial Pb trend and partial Pb loss are viable processes that affected the monazite as most of the monazite experienced dissolution–precipitation processes at temperatures close to or higher than 800 °C (e.g., Kamber and Moorbath 1998; Cherniak et al. 2004; Varga et al. 2020). Such disequilibrium is confirmed by decreasing ²⁰⁶Pb/²³⁸U dates with increasing ThO₂ content (Fig. 3). The sensitivity of the discordance of ²⁰⁷Pb/²³⁵U versus ²⁰⁶Pb/²³⁸U to the Pb-mobilization suggests that both the free- or anchored-regression in the Tera-Wasserburg space is indicative of initial Pb overcorrection. However, with low U-content, the ²⁰⁸Pb/²³²Th and ²⁰⁶Pb/²³⁸U systems are less sensitive to the initial Pb contamination and subsequent mobilization and provide the most robust method of assessing the crystallization age of the monazite (e.g., Barnes et al. 2022). For these reasons, the Th–U–Pb concordia method is assumed to account for the most reliable crystallization ages of the monazite domains.

The two compositionally and texturally distinct monazite groups correlate with different populations of dates (Fig. 4). The moderate-Y domains of group I monazite (samples M92, MP3 and M102) display similar U–Pb lower intercepts and U–Th–Pb (²⁰⁶Pb/²³⁸U versus ²⁰⁸Pb/²³²Th) concordant dates at ~22 Ma, locally preserving relicts with pre-Alpine data associated with the domains with the lowest Y contents (Figs. 3a–c, 4). The domains with moderate-Y and HREE contents are consistent with growth associated with retrogressive garnet breakdown or cooling in the absence of garnet. As the ~22 Ma data cluster recurs in samples either with (MP3 and M102) or without (M92) evidence of partial melting, these data provide no clues on the timing of the melt crystallization. Moreover, widespread ²⁰⁸Pb/²³²Th ages (~22 to 27 Ma) correlate with Th/U increases toward the rims and likely reflect fluid-driven element mobilization during the progressive cooling from supra-solidus to greenschist-facies conditions (Bosse et al. 2009; Seydoux-Guillaume et al. 2012). This suggests that all these rocks (migmatitic or not) follow a common retrogressive path involving garnet breakdown and rehydration that resulted in the most recent monazite growth at ~22 Ma.

The records of group II monazites (samples M69 and M119) are completely different (Figs. 3d–h, 4). Monazite relicts within the allanite–clinozoisite coronas display compelling melt-interaction textures, extremely low contents of Y, Sm, Gd and Dy, and much older ²⁰⁶Pb/²³⁸U versus ²⁰⁸Pb/²³²Th dates of 37–38 Ma. The Y- and Th-depleted monazite population is of particular importance, as marks the end of dissolution processes by the copying or partial inheritance of the precursor grain ages during coupled the ‘smearing’ effect caused (e.g., Grand'Homme et al. 2016; Varga et al. 2020). Therefore, it can be used to constrain the last prograde growth stage of monazite under subsolidus conditions. Noticeably, the rims and fracture zones with higher Th/U ratios record mostly younger ²⁰⁶Pb/²³⁸U versus ²⁰⁸Pb/²³²Th ages, suggesting a relatively late fluid influx that likely induced element mobilization. Consistently, the ²⁰⁸Pb/²³²Th data and the lower intercept date in the Tera–Wasserburg space provide younger ages (reaching ~35 Ma) than U–Th–Pb concordia method.

Because of the absence of the typical post-melting high-Y overgrowths, the monazite record cannot capture the exact timing of the final melt crystallization. Allanite–clinozoisite coronas around monazite are evidence of further interactions with fluids yielding to the replacement of monazite under decreasing P–T conditions (Hentschel et al. 2020; Corvò et al. 2021). However, even these post-melt minerals cannot be used to constrain the timing or conditions of partial melting. An attempt to date allanite following the method of Gregory et al. (2012) provides a statistically insignificant date of 20 ± 41 Ma (MSWD = 0.18; n = 12). Moreover, allanite-to-monazite reactions may occur in a wide P‒T space, between greenschist- and amphibolite-facies conditions (ca. 500–700 °C), depending on the bulk composition, fluid presence and pressure (Spear 2010). Therefore, the allanite–monazite reaction cannot be linked to a precise P‒T space in a straightforward way (Boston et al. 2017).

Zircon age interpretation

The chemical correlation between ages and REE patterns is also commonly used for metamorphic zircon (e.g., Rubatto 2002, 2017; Kohn and Kelly 2018). In general, zircon growth in equilibrium with plagioclase should display a well-pronounced negative Eu anomaly, whereas overall flattening of the M-HREE pattern is expected for zircon grown in equilibrium with garnet (Fig. S2). Additionally, continuous melting and melt loss during protracted supra-solidus conditions result in increasingly negative Eu anomalies (Harley and Nandakumar 2014). Nevertheless, these patterns may be altered by competing chemical reactions under both high- and low-T conditions (Kohn et al. 2015; Kohn and Kelly 2018).

In the Cima di Gagnone metapelites, zircon with variable textures and chemistry correlate with different dated populations (Figs. 5, 6). Zircon crystals from M92 samples display only pre-Alpine concordant U–Pb dates (484 ± 13–422 ± 9 Ma) and only two ²⁰⁶Pb/²³⁸U ages around 33–32 Ma are obtained from two thin bright rims (Figs. 5a–c, 6). The quite steep REE pattern is indicative of magmatic growth, but the narrow dimension of the rims hampers the characterization of the metamorphic overgrowths (Figs. 5a–c, S2). Nevertheless, thin rims are typical of zircons that remain under subsolidus conditions during the metamorphic path (Vorhies et al. 2013; Kohn and Kelly 2018; Tagliaferri et al. 2023).

Zircons from samples M102 and MP3 display pre-Variscan cores (> 500 Ma) and magmatic overgrowths, truncated by thick metamorphic rims with U–Pb dates of 31.4 ± 0.8 and 32 ± 0.6 Ma, respectively (Fig. 6). The mixed Variscan to early alpine dates are characterized by pale patchy textures between clear magmatic overgrowths and rims (Figs. 5d–g, 6). The decreasing HREE contents and the flattening of the slope from magmatic overgrowths toward the metamorphic rims, suggest that the metamorphic rims have grown in equilibrium with the garnet. Both samples preserve clear evidence of partial melting in the rock assemblage (Corvò et al. 2021) and Ti-in-zircon thermometry indicates a minimum temperature of 715 ± 15 °C (Table S5), which is consistent with the presence of melt. Accordingly, the thick rims (10–30 μm) growing on older zircon nuclei are expected to form during cooling and crystallization after partial dissolution of the inherited cores. Therefore, they can be used for dating the onset of cooling before the crystallization of plagioclase or during reduced conditions (e.g., Álvarez-Valero and Kriegsman 2010).

The geochronological patterns of samples M92, M102 and MP3 match the results from the recent geochronological investigations documented in Tagliaferri et al. (2023). Here, several hundred of zircon grains from the Cima Lunga metasedimentary rocks display a main pre-Variscan signal, while the Alpine population at 33–31 Ma has been found only in the syn-tectonic migmatites close to the nappe boundary or to the ultramafic lenses.

Zircons from sample M119 preserve a significantly different geochronological record. Zircon inclusions in garnet, partially or completely shielded by the host, record pre-Alpine dates in the cores mantled by Alpine overgrowths dated between 44.8 ± 4.3 Ma and 37.1 ± 4.7 Ma. These dates likely encompass progressive garnet growth (Figs. 5j–l, 6). In the matrix, very large zircons completely lack pre-Alpine domains and display complex sector-zoned cores mantled by a relatively thick (5–30 μm) outer rim. Multiple, internal domains record Alpine ages between ~38 and ~35 Ma, which correlate with HT conditions (Ti-in-zircon value of 750 ± 22 °C), whereas the outer rims preserve slightly younger ages (~35–33 Ma) and lower temperatures of ~700 °C (Table S5). The highly variable contents of Y, U and REEs associated with general depletions in Pb and Th and low Th/U, suggest that zircon interacts with a local melt that continuously changes its composition during the dissolution–precipitation process (e.g., Kusiak et al. 2013; Peterman et al. 2019; Varga et al. 2020).

This chemical signature is consistent with zircon harboring large and multiple overgrowths typical of leucosomes (Álvarez‐Valero and Kriegsman 2010; Kohn and Kelly 2018). The absence of pre-Alpine ages in the matrix crystals indicates that detrital zircons completely dissolved during the migmatization of the rock, or they reset due to the reactions with the alkaline fluids/melts (Schenker et al. 2018). Therefore, we argue that the age of ~35 Ma represents the time of post-melt crystallization, while the interval ~38 to 35 Ma could represent the maximum duration of supra-solidus conditions. Finally, the slightly younger ages reaching ~31 Ma from the outer rims likely reflect further dissolution–precipitation processes at decreasing temperatures.

Timing of metamorphism and partial melting

The study of zircon and monazite highlights a complex metamorphic history recorded by the Cima di Gagnone rocks. Each sample preserves a different geochronological record, which may correlate to either a different segment of the P‒T path or mixed ages without geological meaning.

Zircon and monazite U–Th–Pb data from the metapelite sample closest to the ultramafic bodies (sample M119) allow us to constrain the melting stage between ~38 and ~35 Ma at a minimum temperature of ~750 °C, while between ~33 and 31 Ma further dissolution–precipitation processes occurred at lower temperature of ~700 °C (Figs. 7, 8a). While zircon in the matrix crystallized during or after the main melting event, the ~44 to 37 Ma age of zircons included in garnets coincide with or predate garnet growth with respect to the melting stage. The monazite/garnet REE pattern and partition coefficients suggest chemical equilibrium between the monazite relicts and garnet cores (Fig. S1). The mineral assemblage of the garnet core inclusions (phengite + biotite + rutile + quartz + plagioclase), geothermobarometry and thermodynamic modeling indicate P‒T values of ~1.4 to 1.7 GPa and 780–850 °C, respectively (Corvò et al. 2021). These conditions represent a reliable proxy of the HT melting stage (Fig. 8a). However, this conclusion does not exclude the possibility that the sample experienced previously (~43 to 38 Ma) higher P conditions up to 2.5–2.7 GPa (Corvò et al. 2021; Piccoli et al. 2022). Cooling at T < 700 °C is characterized by fluid-mediated metasomatic reactions (Corvò et al. 2021) that locally resulted in the formation of allanite–clinozoisite coronas; however, the timing of this event cannot be precisely constrained. The other sample (M69) with similar corona structures records the same main melting stage at ~37 Ma. The slightly lower P‒T conditions inferred from this sample (1.2–1.4 GPa and 720–750 °C; Fig. 8a) may reflect either a retrogressive stage, or the local development of lower T conditions (Corvò et al. 2021).

Fig. 8

a Geochronological results correlated with the compilation of P–T fields of (ultra)mafic rocks (green), metapelites (orange) and migmatitic metapelites (blue). P‒T estimations from samples M92, M102, MP3 and M119 are from Corvò et al. (2021), from sample D2 from Piccoli et al. (2022). Other estimates are from: (1) Nimis and Trommsdorff (2001) and Scambelluri et al. (2014); (2) Brouwer et al. (2005); (3) Grond et al. (1995); (4) Heinrich (1986). Geochronological data are from this study, and (5) Gebauer (1996, 1999) and 6 Becker (1993). The black line indicates the retrograde path—common to all samples—toward greenschists facies. b Hypothetical P‒T–t paths projected on the estimated P‒T fields for all the Cima Lunga samples. The two alternative exhumation trajectories assume a single and common P‒T–t path for all samples. Red line considers a nearly isothermal decompression from ~3 to 1 GPa (Heinrich 1982; Grond et al. 1995; Pfiffner 1999; Nimis and Trommsdorff 2001), fitting the largest number of P‒T fields. Purple line shows an exhumation path characterized by nearly isobaric cooling at HP conditions, followed by isothermal decompression up to ca- 1 GPa, as suggested by Piccoli et al. 2022. Dashed line indicates the melt curve representative for semipelitic composition (e.g., MP3; Corvò et al. 2021)

The other migmatitic samples (M102 and MP3) do not record the ~38 to 35 Ma melting stage in either zircon or monazite. Zircon preserve pre-Variscan magmatic overgrowths and pale zircon zones showing young, scattered ages down to the Cretaceous. According to the unclear texture around these spot analyses, the scattered Permian and Cretaceous ages are interpreted as mixing values between the old magmatic cores and the metamorphic rims. Thinner zircon rims register less pronounced dissolution–precipitation textures developed at a minimum temperature of 715 ± 15 °C and dated to ~33 to 31 Ma. These conditions are consistent with the P‒T estimates derived from the mineral assemblages of the leucosomes from the large migmatitic patches, from where samples record temperatures between 760 and 700 °C and variable pressures between 1.7 and 0.8 GPa (Fig. 8a; Corvò et al. 2021). This 33–31 Ma migmatitic phase is also well recorded throughout the Cima Lunga unit, particularly along the upper tectonic contact with the Maggia unit (Tagliaferri et al. 2023).

The monazite record in these samples (M102 and MP3) indicates that some pre-Alpine cores are preserved, although a large number of the grains record a retrogressive path marked by a moderate-Y content and ages between ~27 and ~22 Ma associated with garnet breakdown and fluid-rock interactions (Fig. 8a). The final age of ~22 Ma likely represents late overgrowth of monazite during cooling to lower amphibolite-facies conditions.

Finally, the sample without evidence of partial melting structures (M92) registers very limited traces of the ~33 to 31 Ma event in very thin zircon rims, which likely remained under subsolidus conditions during the entire Alpine metamorphic path, as inferred from the P‒T estimates (P < 1.1 GPa and T < 700 °C; Fig. 8a). The monazite data are consistent with those obtained from samples M102 andMP3, confirming that the most recent monazite grew onto pre-Alpine cores at ~22 Ma. This age likely also corresponds to the substitution of monazite with allanite, less than ca. 600 °C (Spear 2010).

Melt/fluid–rock interactions

The heterogeneous P‒T–t fields inferred from each sample are influenced by the interactions between the rocks and metamorphic fluids or melts. A detailed study on the geochemical interactions between ultramafic rocks and metapelites demonstrated long history of episodic fluid-mediated exchange of mobile elements between the rock types, locally enhancing partial melting (Pfeifer 1981, 1987; Heinrich 1982; Früh-Green 1987; Pfiffner 1999; Scambelluri et al. 2014, 2015; Cannaò et al. 2015; Corvò et al. 2021). In particular, the chemical signature recorded in the dated migmatitic samples is due to the interactions with a fluid sourced from the contiguous hydrated ultramafic rocks enriched in CaO, Al₂O₃, FeO and MgO, but undersaturated in SiO₂ (Pfeifer 1981; Corvò et al. 2021). Recent studies (Corvò et al. 2021; Piccoli et al. 2022) revealed that some of these interactions occurred under HT conditions (T > 700 °C). Migmatitic patches occur as patches located in dilatant sites, such as pressure shadows or fractures of the lenses, supporting for a process of dynamic porosity increase, favoring a pressure gradient that could have localized an outwards flux of fluids from the ultramafic lenses. The combination of high temperature with the circulation of free fluids suggests the activation of fluid-fluxed melting, which is a recognized viable mechanism to explain the production of migmatites and even granitic magmas (e.g., Weinberg and Hasalová 2015; Tafur and Diener 2020; Casini et al. 2023). Ambiguity remains as to whether the HT conditions affected all the metasedimentary rocks during a common metamorphic history or only a few spots located around the lenses, as a result of a local process.

According to the dating results, small portions of the metasedimentary rocks record migmatization that developed between ~38 and ~30 Ma, whereas colder fluid–rock interactions (< 700 °C) lasted until ~22 Ma, resulting in lower amphibolite facies conditions. There is a robust agreement between the preserved mineral assemblages and microstructures (as well as between the zircon and monazite records) and the calculated melt volume in all samples (Fig. 7). The significant melt production registered by the migmatitic metapelites relies on the efficacy of these dehydration processes during nearly isothermal decompression from HP at T > 700 °C. Under these conditions, the release of H₂O-rich fluids may result from the breakdown of micas, following the generalized reaction (Heinrich 1982):

$${\text{phengite}} + {\text{garnet}} \pm {\text{paragonite}} = {\text{muscovite}} + {\text{biotite}} + {\text{plagioclase}} + {\text{quartz}} + {\text{H}}_{2} {\text{O}}.$$

(1)

The seminal paper of Heinrich (1982) revealed the extensive amphibolite-facies retrogression of the Cima Lunga and Adula (ultra)mafic lenses as consequence of metapelite dehydration, resulting in closed-system behavior at the scale of a tectonic unit. According to this model, metapelites, thus, represent the source of fluids triggering the rehydration of otherwise anhydrous (ultra)mafic rocks, which acted as fluid sinks. On the other hand, Corvò et al. (2021) provided evidence of how the bulk composition and the mineral assemblage of studied samples result from the fluids sourced from the hydrated ultramafic lenses (amphibole- and chlorite-rich), which are enriched in CaO, Al₂O₃, FeO, and MgO components and depleted in SiO₂ during fluid-mediated retrogression. Textural, chemical and geochronological records of monazite and zircon further support such a picture of strong fluid–rock interactions localized in the halos of the (ultra)mafic lenses. Such a context of sustained HT fluid-rock interactions is the worst condition for preserving the HP mineral assemblage. In contrast, metapelites following decompression at T < 700 °C would have recorded only occasional, limited, fluid-rock interactions without producing detectable amount of melt, likely preserving relicts of their HP mineral assemblage (Fig. 7). This observation agrees with the metamorphic record preserved by metapelite in other eclogite-type localities, which record near-isothermal decompression from 2 to 1 GPa at 670–690 °C under fluid-absent conditions (e.g., Saualpe–Koralpe, Eastern Alps, Austria; Schorn 2017).

Taken together, the literature and the results obtained in this study indicate that the decompressional path of the Cima di Gagnone–Cima Lunga unit was characterized by the periodic reversal of fluid flow polarity across the interface between metapelite and ultramafic rocks. Counterintuitively, such an interface preserves the record of the HP assemblage although this should be easily removed by the enhanced fluid action.

P‒T–t paths

The process of assembling the heterogeneous P‒T–t fields inferred from each sample within a coherent geodynamic model requires the integration of complex structural, geochronological and metamorphic records in relation to partial melting and metasomatic processes. Provided that the exhumation rate is fast enough to prevent excessive conductive heat loss, the simplest initial model to start with is exhumation by nearly isothermal decompression (red line in Fig. 8b). Therefore, exhumation occurred at T > 700 °C for approximately 9–4 Ma (considering the time interval of 40–31 Ma). The absence of extensive migmatization in the Cima Lunga unit implies that the metapelites remained mostly anhydrous (e.g., without a significant amount of free water), so that melting would have required a temperature > 950 °C (Pfiffner 1999; Hermann and Green 2001; Spandler and O’Neill 2010). The presence of migmatites should therefore be justified through injections of exotic fluids triggering localized fluid-fluxed melting, leaving the rest of the unit under subsolidus conditions. However, this model contrasts with the mineral assemblages preserved in metapelites, mafic and ultramafic rocks, which widely preserve stable hydrated phases stable from pre- to peak and retrograde conditions (Pfeifer 1981, 1987; Heinrich 1982; Früh-Green 1987; Pfiffner 1999; Nimis and Trommsdorff 2001; Brouwer et al. 2005; Scambelluri et al. 2014, 2015; Cannaò et al. 2015; Corvò et al. 2021). Moreover, thermodynamic modeling confirmed that HT decompression of metapelites would have produced significant volume of melt in all the mineral assemblages (Fig. 7), and not only around ultramafic bodies or within major shear zones. Therefore, the model of simple nearly isothermal exhumation must be discarded due to the scarcity of migmatites and relicts of HT assemblages.

Alternatively, a counterclockwise trajectory has been alternatively proposed to describe the exhumation path of the Cima Lunga unit (light purple in Fig. 8b; Piccoli et al. 2022). These authors proposed that a nearly isobaric cooling stage soon after the P‒T peak could have allowed isothermal decompression at much lower temperatures, preventing massive melting during exhumation (Fig. 8b). However, this model does not account for the P‒T conditions inferred from most of the rocks of the Cima Lunga unit (Evans et al. 1979; Heinrich 1982; Früh-Green 1987; Grond et al. 1995; Pfiffner and Trommsdorff 1998; Pfiffner 1999; Brouwer et al. 2005; Cannaò et al. 2015; Corvò et al. 2021). In particular, this model fails to explain the samples that record HT at pressures well below the metamorphic peak, as well as the compelling field, petrological and structural evidence of migmatites chronologically constrained during the Alpine exhumation phase from 35 to 31 Ma (Fig. 2; see also Corvò et al. 2021; Tagliaferri et al. 2023).

The apparent difficulty in reconciling the rock heterogeneity in a single coherent path seems to favor another geodynamic model in which the various samples evolve independently of each other prior to sharing a common exhumation path after reaching subsolidus amphibolite-facies conditions. In this view, the amphibolite-facies metamorphism would represent the regional conditions experienced by the entire Cima Lunga unit, whereas the higher-temperature conditions should be produced by a local mechanism across the metapelite–ultramafic rocks interface. This model cautions that differences in P‒T record may track different rheological responses of compositionally heterogeneous rocks deforming at the same depth.

Tectonic implications

All studies in the Cima Lunga unit indicate that P‒T peak conditions were attained between ~40 and ~35 Ma (Figs. 8a, 9; e.g., Becker 1993; Gebauer 1996; Brouwer et al. 2005; this study), in agreement with the other (U)HP occurrences of the Central Alps (Fig. 9; Alpe Arami, Monte Duria, Capoli and the Adula nappe, e.g., Hermann et al. 2006; Herwartz et al. 2011). However, U–Pb zircon and allanite ages between ~33 and ~30 Ma are instead interpreted as thermal perturbations associated with nappe stacking (Tagliaferri et al. 2023) or related to post-nappe stacking thermal re-equilibration and Peri-Adriatic magmatism (Gebauer 1996; Liati et al. 2009; Gregory et al. 2012; Cavargna-Sani et al. 2014; Boston et al. 2017). Monazite ages of ~22 Ma are recorded for the entire Lepontine dome (Köppel and Grünenfelder 1979; Janots et al. 2009; Boston et al. 2017), suggesting a common post-Barrovian cooling history (Tagliaferri et al. 2023). The new dataset presented in this study confirms the above findings but reveals an important, although localized, melting stage at ~38 to 35 Ma at high temperatures (~750 to 850 °C) and intermediate pressures (~1.7 to 1.5 GPa) between the inferred pressure peak (2.5–3 GPa) and subsequent decompression to amphibolite-facies conditions (0.8–1 GPa). However, the record of such a melting stage is limited to metapelites close to only one ultramafic body, whereas the other migmatites that developed around the ultramafic lenses and the upper thrust zone are dated at ~33 to 30 Ma with T values around 750–700 °C (this study and Tagliaferri et al. 2023). Noticeably, the rest of the unit (i.e., the vast majority) lacks any records of partial melting. The significance of these extremely localized HT events remains an open question, challenging its tectonic significance. In the other UHP ultramafic lenses from the Central Alps (Monte Duria, Alpe Arami and Capoli), HT conditions are described as preserved from the P‒T peak along mostly isothermal decompression paths, prior to cooling below ca. 1 GPa (Fig. 9; Brouwer et al. 2005; Hermann et al. 2006; Tumiati et al. 2018; Pellegrino et al. 2020). However, the geochronological record of the metapelites surrounding these lenses was strongly influenced by the 33–30 Ma melting event, which was more intense close to the Southern Steep Belt (Fig. 1b; e.g., von Blanckenburg and Davies 1995; Burri et al. 2005; Gregory et al. 2012). Moreover, the Cima Lunga unit experienced higher temperatures than did the other (U)HP units of the Western Alps (e.g., Dora-Maira, Lago di Cignana, Zermatt-Sass, Monviso, Gran Paradiso and Monte Rosa) characterized by colder exhumation paths at temperatures mostly below 700 °C (Fig. 9; e.g., Compagnoni et al. 1995; Agard et al. 2001; Rubatto and Hermann 2001; Angiboust et al. 2009; Manzotti et al. 2015, 2018, 2022; Groppo et al. 2009; Locatelli et al. 2018; Luoni et al. 2021). Accordingly, these relatively colder units do not show evidence of partial melting. The Cima Lunga unit therefore represents a unique opportunity to study the behaviour of metapelitic rocks under (U)HP–HT conditions.

Fig. 9

Compilation of P‒T estimates (black squares) from (U)HP rocks of Central Alps, including: AA – Alpe Arami (AA-per Nimis and Trommsdorff 2001; AA-ecl Brouwer et al. 2005); Bo -Borgo (Tumiati et al. 2018); Cap – Capoli (Brouwer et al. 2005); CdG – Cima di Gagnone peridotite (CdG1 Pfiffner 1999; CdG2 Nimis and Trommsdorff 2001; CdG3 Scambelluri et al. 2014); CL-ecl – eclogite from Cima Lunga unit (Brouwer et al. 2005); Cl-met – metasedimentary rocks from Cima Lunga Unit (Piccoli et al. 2022); Du – Monte Duria (Du1 Nimis and Trommsdorff 2001; Du2 Hermann et al. 2006; Du3 Tumiati et al. 2018); Gor – Gorduno (Brouwer et al. 2005); SAd – Southern Adula (Dale and Holland 2003); SSB-mig – migmatite of Southern Step Belt (Burri et al. 2005); Tres – Trescolmen (Meyre et al. 1999). PT fields from Corvò et al. (2021) (M92, M69; MP3; M119; M102) and Piccoli et al. (2022) (D2) are also projected. Colored lines indicate the PTt paths proposed for representative (U)HP units of Central: AA – Alpe Arami (green; Nimis and Trommsdorff 2001; Brouwer et al. 2005) CdG – Cima di Gagnone (red; Nimis and Trommsdorff 2001); Monte Duria (blue; Tumiati et al. 2018). Note that none of the paths fit the wide spread of the P‒T fields of the related unit. PT paths from the Western Alps (light purple) are also plotted: DM-BIU – Dora Maira Massif, Brossasco-Isarca Unit (Hermann 2003; Groppo et al. 2016); GP – Gran Paradiso unit (Manzotti et al. 2015, 2018); LCU-ZS – Lago di Cignana Unit-Zermatt Saas Zone (Angiboust et al. 2009; Groppo et al. 2016). Geochronological data are from this study, Gebauer (1996, 1999), Becker (1993), Hermann et al. (2006), and Herwartz et al. (2011). Dashed lines indicate the melt curves (wet melting¹ and wet melting²) representative for a generic metapelitic composition (wet melting¹; Spandler et al. 2007) or specific for the composition of sample MP3 (wet melting²; Corvò et al. 2021)

To date, three interpretations have been proposed to explain the high variability of P‒T–t paths determined in the Cima Lunga unit and the similar Adula nappe: (i) Alpine channel flow yielding to a tectonic mélange (Trommsdorff 1990; Engi et al. 2001; Berger et al. 2005; Cannaò et al. 2015; Scambelluri et al. 2015); (ii) heterogeneous, compositionally controlled, mineralogical re-equilibration within a coherent tectonic unit that experienced a uniform prograde–retrograde path (Heinrich 1982; Nagel 2008; Herwartz et al. 2011; Sandmann et al. 2014; Piccoli et al. 2022); and (iii) local variations in both pressure and temperature conditions, or local changes in the mineral stability field, due to non-hydrostatic stress (Pleuger and Podladchikov 2014; Corvò et al. 2021; Tagliaferri et al. 2023). The pre-Alpine origin of the Cima Lunga sedimentary mélange (Tagliaferri et al. 2023), the structural coherence of the Cima Lunga unit and the coupled—although heterogeneous—strain evolution experienced by metapelites and ultramafic rocks argue against the model of an Alpine tectonic mélange (Maino et al. 2021; Tagliaferri et al. 2023). On the other hand, the HT conditions of 700–850 °C recorded during exhumation by a small but significant number of samples cannot be extrapolated to the whole unit, as they are incompatible with both the limited volume of melts and the extreme localization of migmatites. In contrast, the HP relicts preserved in the area are apparently affected by the strongest fluid– and melt–rock interactions. This contradicts the assumption that anhydrous conditions are needed to explain the absence of massive melting (Sandmann et al. 2014; Piccoli et al. 2022). All these observations suggest that irreconcilable metamorphic records reflect local variations of metamorphic conditions, rather than differential preservation of near-peak P‒T conditions.

Localized thermal perturbations over a few meters may be explained by the rapid advection of exotic hot fluids or by shear heating (e.g., Gottardi et al. 2011; Mako and Caddick 2018; Maino et al. 2015; 2020; Casini et al. 2021, 2023; Tagliaferri et al. 2023). However, the episodic fluid-mediated chemical exchange between ultramafic and the adjacent metasedimentary rocks since prograde to retrograde metamorphic conditions (Pfeifer 1981, 1987; Heinrich 1982; Scambelluri et al. 2015; Cannaò et al. 2015; Corvò et al. 2021) suggests a local source of fluids sourcing from the ultramafic rock, excluding deep-sourced external inputs. Hence, the structure and composition of the Cima Lunga unit support localized and episodic shear heating within the ultramafic rocks and at their interface with metapelites. The efficacy of this mechanism is greatly enhanced by high shear strain (γ) and large viscosity contrast (m) (Gerya 2010; Burg and Gerya 2005; Casini and Maino 2018; Maino et al. 2020; Vaughan-Hammon et al. 2022; Tagliaferri et al. 2023), similar to those calculated for the metapelite–ultramafic pairs of the Cima Lunga unit (γ > 14.5; m up to 9; Maino et al. 2021). Moreover, transient higher-temperature pulses (up to 100 °C) may be produced by coseismic slip (Maino et al. 2020) possibly developed within the cataclastic zones displayed by the ultramafic lenses, locally filled by melt (Fig. 2c). It should be noted that the structural configuration enhancing shear heating is also prone to build up tectonic overpressure (e.g., positive deviation from lithostatic pressure) in the strongest ultramafic bodies (e.g., see Gerya 2015 for a review). According to this model, strong and relatively rigid inclusions set within a weaker matrix may concentrate significant deviatoric stress that can double the lithostatic pressure (Mancktelow 1993, 2008; Schmid and Podladchikov 2003; Moulas et al. 2014; Pleuger and Podladchikov 2014; Gerya 2015; Schenker et al. 2015; Chen et al. 2017; Casini and Maino 2018; Luisier et al. 2019). Therefore, at least some of the pressure variation in the various rocks of the Cima Lunga unit could be explained as an effect of tectonic overpressure.

In this view, the heterogeneous P‒T evolution of the (ultra)mafic and metapelite pairs is accounted for by the variable effect of stress along the rheological interfaces in association with the fluid interactions during the tectonic evolution of the Cima Lunga unit (Fig. 10). During Alpine evolution, the Cima Lunga unit acted as a major shear zone sandwiched between the Maggia and the Simano nappe (Tagliaferri et al. 2023). The 38 Ma age revealed by garnet Sm–Nd and zircon U–Pb from ultramafic rocks (Becker 1993; Gebauer 1999) and zircon and monazite U–Pb from metapelites (this study) may indicate the effectiveness of shear heating to enhance HT conditions around the ultramafic lenses (Fig. 10a, d). Transient pressure and temperature deviations likely developed along the rheological boundaries and along the main nappe thrust where the deformation-induced thermal anomaly is apparently sustained until ~31 Ma (Fig. 10b, e–f). After this stage, thermal relaxation associated with the thickened crust is manifested through a slow cooling of the unit, following the N–S regional gradient (Fig. 10c, g).

Fig. 10

Geodynamic evolution of the Cima Lunga unit within the Lepontine nappes (modified from Tagliaferri et al. 2023). References for depth- and T-estimates, and literature ages are mentioned in the main text. a Beginning of overthrusting during the coherent exhumation of the Maggia over the Simano nappes. The position of the Cima Lunga unit corresponds to the main shear zone. Shear senses and related local stress fields are indicated in the dashed squares (OP: overpressure, LP: lithostatic pressure). (Ultra)mafics are indicated with green ellipses. b Internal structure of the Lepontine nappe stack at the regional peak Barrovian conditions at 31 Ma. The white dashed lines are the indicative isotherms. In the box: section of the eastern sector, after Galli et al. (2012). AA: Alpe Arami; MD: Monte Duria. c Overall thermal re-equilibration of the Lepontine dome. The relaxation of the isotherms (cfr. the white dashed line with the line in transparency) corresponds to local heating. In the box: section of the eastern sector, the age of Novate granite is after Liati et al. (2000). d–g Sketch at outcrop scale displaying the evolution of the (ultra)mafic–metapelite pairs during the Alpine deformation and metamorphism. Evolution of melt patches around the (ultra)mafic lenses marks the HT decompressive stages captured by the zircon and monazite geochronometers

Conclusions

The pre-Alpine metasedimentary rocks of the Cima Lunga unit, Central Alps, experienced distinct P‒T conditions associated with different times of Alpine metamorphism. Zircon U–Pb dating of the metapelites with a peak pressure < 1.1 GPa and temperature < 700 °C, indicates Alpine metamorphism, as only pre-Variscan or Variscan ages are recorded. Rare thin zircon rims (< 10 μm) register U–Pb ages of ~33 to 31 Ma, which is consistent with the occurrence of Barrovian thermal perturbation during the vanishing stage of decompression and nappe emplacement. Monazite age of ~22 Ma marks instead the cooling at lower amphibolite-facies conditions during the Barrovian stage.

In contrast, at the interface with the ultramafic lenses, metapelites developed higher conditions of 1.3–2.7 GPa and 700–850 °C, at supra-solidus conditions. Metasomatic reactions are linked to the episodic exchange of fluids along the entire P‒T path and locally promoted spotted melting stages recorded by zircon and monazite growth at ~38 to 35 Ma and 33–31 Ma, respectively. Quantification of fluid availability during decompression shows that melt processes are coherently expected at T > 700 °C. The mechanisms by which such coexisting, although different, metamorphic and geochronological records developed are enigmatic. The tectonic mélange or heterogeneous re-equilibration of a coherent tectonic unit that experienced uniform UHP and HT conditions does not account for the structural coherence of the unit and/or the localization of high-T conditions and melting processes in certain domains of metapelites during exhumation. Distinct melting episodes at 38–35 and 33–31 Ma are interpreted in relation to localized fluid flux at the interfaces between metapelite and ultramafic rocks, possibly enhanced by transient shear heating. Our research emphasizes that significant P‒T variations at the local scale may result from the effects of non-hydrostatic stress and shear heating coupled with the presence/absence of fluids, ultimately influencing the evolution of metamorphic systems.