Constraints on the preservation of proxy data in carbonate archives – lessons from a marine limestone to marble transect, Latemar, Italy

This work evaluates an exceptionally complex natural laboratory, the Middle Triassic Latemar isolated platform in the northern Italian Dolomite Mountains and explores spatial and temporal gradients in processes and products related to contact metamorphism, dolomitization and dedolomitization of marine limestones. The relation between petrographic change and re‐equilibration of geochemical proxy data is evaluated from the perspective of carbonate‐archive research. Hydrothermal dolomitization of the limestone units is triggered by dykes and associated hydrothermal fluids radiating from the nearby Predazzo Intrusion. Detailed petrography, fluid inclusion analysis, δ13C and δ18O data and 87Sr/86Sr isotope ratios shed light on the extreme textural and geochemical complexity. Metamorphic and diagenetic patterns include: (i) peak‐metamorphic and retrograde‐metamorphic phases including three dolomite marbles, two dedolomite marbles, brucite, magnesium silicates and late‐stage meteoric/vadose cement at the contact aureole; (ii) four spatially defined episodes of dolomitization, authigenic quartz, low magnesium calcite and late‐stage meteoric cement at the Latemar isolated platform; and (iii) kilometre‐scale gradients in δ13C values from the contact aureole towards the platform interior. Results shown here are relevant for two reasons: first, the spatial analysis of alteration products ranging from high‐grade metamorphic overprint of marbles at temperatures of 700°C in the contact aureole to moderately altered limestones in the platform interior at temperatures <100°C, allows the observation of processes that commonly occur along vertical (prograde) gradients from shallow burial to metamorphism at depths >20 km. Second, under rock‐buffered conditions, and irrespective of metamorphic to diagenetic fluid−rock interactions, both marbles, and low‐temperature hydrothermal dolomites have conservative marine δ13C and δ18O values. The fact that metamorphism and hydrothermal dolomitization of precursor limestones and early diagenetic dolostones did not per se reset environmental proxy data is of interest for those concerned with carbonate archive research in Earth’s deep time.


INTRODUCTION
Carbonate archive data represent a snapshot of the evolution of Earth's surface environments, recording complex interactions of equilibrium and non-equilibrium processes. However, carbonate archives are susceptible to postdepositional alteration (e.g. Bathurst, 1975;Swart, 2015). The challenge is to separate a meaningful environmental signal from one that is modified by diagenetic or metamorphic alteration, and hence to reliably assess the integrity and robustness of the extracted data as indicators of past depositional environments. Dolomitization, i.e. calcite to dolomite neomorphism, is prominent among the complex array of diagenetic processes, and may result in partial to complete resetting of primary geochemical values Mueller et al., 2020). Metamorphic alteration of archive data by regional or contact metamorphism is documented in numerous examples worldwide (Wada et al., 1998;Bowman et al., 2009;Boomeri et al., 2010). These processes are of fundamental relevance in deep-time climate dynamics research, specifically concerning the Palaeoproterozoic and Archean, which are characterized by abundant, often overprinted dolomite archives (Veizer et al., 1989;Brady et al., 1998;Bau et al., 1999;Franchi & Abebe, 2020).
Here, it is argued that contact metamorphism is relevant, albeit underexplored, for those concerned with palaeoenvironmental data from dolomitized carbonate archives. The rationale is, that contact metamorphic aureoles of magmatic bodies in limestones offer unique and accessible windows into processes and features that otherwise take place along burial to metamorphic gradients from near-surface down to 20 km or more of burial depth.
The Anisian-Ladinian Latemar isolated platform and the neighbouring Predazzo magmatic intrusion in the northern Italian Dolomite Mountains, form an ideal case example in this sense. At Latemar, 74% of the calcitic mineralogy is preserved (Jacquemyn et al., 2014), allowing for the investigation of a marine limestone to marble transect. The Latemar isolated platform is one of the most intensively studied carbonate edifices worldwide, with previous work focusing (among other topics) on bedding patterns, orbital forcing and cycle patterns, evidence for subaerial exposure, stratigraphy, platform geometry, reservoir analogue properties and dolomitization (see references in Christ et al., 2012; Table 1). Work on the Predazzo Intrusion metamorphic aureole includes that of Ferry et al. (2002), dealing with the direction of fluid flow, and various petrological studies (Castellarin et al., 1982;Sloman, 1989;Marrocchino et al., 2002Marrocchino et al., , 2009Casetta et al., 2018a, b). Jacquemyn et al. (2014Jacquemyn et al. ( , 2015 developed a paragenetic sequence and revealed the exceptionally complex spatial relation between platform carbonates and the Predazzo sub-volcanic intrusion (Castellarin et al., 1982). The intrusion metamorphosed underlying strata and is genetically related to swarms of dykes that cross-cut the platform and formed transport and interaction pathways for dolomitizing fluids. Previous work includes a wide range of dolomitizing fluid temperatures. Based on fluid inclusion thermometry, Wilson et al. (1990) suggested 75 to 220°C, whereas Ferry et al. (2011) proposed temperatures from 40 to 80°C based on carbonate clumped isotope thermometry.
Research presented here combines new evidence based on petrographic (transmitted light and cathodoluminescence microscopy, electron microprobe, microthermometry) and geochemical (d 18 O, d 13 C, 87 Sr/ 86 Sr) analysis with published data from Jacquemyn (2013) and Jacquemyn et al. (2014Jacquemyn et al. ( , 2015. This paper has three aims: (i) to characterize a petrographic and geochemical transect across the complex metamorphic to low-temperature diagenetic reaction pathway of Triassic Latemar carbonates; (ii) to systematically place these data into spatial and temporal context ranging from contact metamorphic (de-)dolomite marbles (700°C; Ferry et al., 2002) to marine limestones exposed to low-temperature diagenetic fluids (<80°C; Ferry et al., 2011); and (iii) to establish tipping points and tools for their recognition along the alteration pathway of a marine carbonate archive exposed to increasing degrees of overprint.
Acknowledging the inherent complexity of carbonate archives and their proxy data, the authors do not claim that the Latemar case example represents a litmus test for any given outcrop throughout Earth´s geological history. In contrast, this contribution aims to provide a framework against which existing and subsequent work can be placed to extract quantitative science, patterns and methods for the community.
Carbonate deposition ceased with the onset of the Predazzo Intrusion in the Late Ladinian (Bosellini, 1984). A Plinian/Vesuvian eruptive phase draped the Latemar with volcanic material and locally caused contact metamorphic overprinting of the carbonates (Fig. 2C). The subsequent emplacement of the Predazzo Intrusion took place in several phases from 238 to 232 Ma (Castellarin et al., 1982;Mundil et al., 1996Mundil et al., , 2010. The Predazzo Intrusion is composed of approximately coeval intrusive bodies arranged in a circular arrangement, 2 km in diameter, and is controlled by faults that formed during caldera collapse . Three magmatic pulses are recorded: two shoshonitic (silica saturated: SiSa; silica undersaturated: SiUSa) and one granitic unit (GU; Fig. 2D to F; Casetta et al., 2018a,b). The relative age of each pulse was determined by cross-cutting relationships (Marrocchino et al., 2009). A contact metamorphic aureole (T ≤ 700°C) composed of dedolomitized marbles formed around the Predazzo Intrusion (Fig. 2C), and locally contains up to 40 vol.% of brucite [Mg(OH) 2 ] and a series of complex silicates (Jacquemyn et al., 2014). This marble unit represents the high-temperature target lithology in this paper.
The intrusion was coupled with the emplacement of numerous north/north-west to south/ south-east oriented dykes cross-cutting the Latemar isolated platform (Riva & Stefani, 2003). These dykes correspond to the three aforementioned magmatic pulses and bear a similar composition, revealing their relative timing (Marrocchino et al., 2009;Casetta et al., 2018a, b). High-permeability brecciated zones along the dyke margins served as fluid pathways that caused partial (26%) and spatially localized dolomitization of the otherwise calcitic Latemar isolated platform ( Fig. 3; Jacquemyn et al., Table 1. Summary of a selected set of important studies of the Latemar carbonate buildup in the past decades. Studies are grouped by their main focus, topics and research fields. Many more studies, especially with regional geological focus, are not included in this table.

Fieldwork and sampling strategy
The following sampling strategy was applied to build a transect from most to least overprinted.
1 The Canzoccoli quarry, located 3 to 4 km to the south-east of the Latemar at 1000 m above sea-level (a.s.l.), serves as a high-temperature (T ≤ 700°C, sensu Ferry et al., 2002) Fig. 2B; Fig. 3D and E), represents an intermediate setting between contact metamorphism and moderate hydrothermal alteration of 'Latemar slope samples'.
3 The Latemar mountain (platform interior) located 5 to 7 km north-west of the intrusion between 2400 and 2600 m a.s.l., exposes the highest occurrence of partially dolomitized Sciliar Formation, and access to the lowtemperature (T < 100°C) end member. 'Latemar interior' samples represent low-temperature hydrothermal dolomitization (Middle Tepee Facies, Lower Cyclic Facies and Lower Tepee Facies; array #3 in Fig. 2B and Fig. 3A to C).
A complete (metamorphic) phase sequence at the contact metamorphic aureole was established along this transect and complemented the sequence from marine diagenetic limestones to dedolomites at the Latemar with calcsilicates in the context of a peak-metamorphic to retrograde diagenetic pathway. All visible rock types, vein types and cement types were sampled and characterized by thin section and geochemical analyses, enabling the assessment of temperaturerelated and fluid-related petrographic and geochemical differences within their spatial and petrographic context. Dolomite replacement fabrics and cement phases were documented in their relation to dykes. To identify timeequivalent facies between Latemar carbonates (Sciliar Formation) and the metamorphosed dolostones at the Predazzo Intrusion contact (Contrin Formation), the basal, early diagenetic dolostones of the Contrin and Werfen (shallow water marls) formations exposed at the Latemar slope were sampled. Note that arrays 2 and 3 are stratigraphically distinct as they cover different parts of the isolated platform. Refer to Appendix S1 for further information on sampling locations and measured sections (Fig. S1).

Petrographic and mineralogical analysis
In total, 119 thin sections were analyzed using polarized (n = 119), cathodoluminescence (n = 41) and scanning electron microscopy (n = 4). Polarized photomicrographs were taken with a Leica DM4500P microscope (Leica Microsystems GmbH, Wetzlar, Germany). The thin sections were sputter-coated with a 15 nm thick gold layer to avoid charging for both cathodoluminescence and scanning electron microscopy. Cathodoluminescence microscopy was performed using a hot cathode (HC1-LM) facility at the Ruhr-University Bochum (Neuser et al., 1996) equipped with a DC73 video camera system (Olympus Corporation, Tokyo, Japan). The electron beam had an acceleration voltage of 14 kV, current density from 5 to 10 μA/mm², and beam current between 0.1 and 0.2 mA. For scanning electron imaging, a high-resolution field emission scanning electron microscope (HR-FESEM) type Merlin Gemini II (Carl Zeiss AG, Jena, Germany) using secondary electrons was used at the Ruhr-University Bochum.
To identify mineral phases, 105 samples were analyzed with a Philips X´Pert MPD Theta-Theta X-ray Diffractometer (Malvern PANalytical, Malvern, UK) using Cu Ka radiation at 45 kV and 40 mA. The scanning range was 0 to 70°(2h), with a step of 0.025°(2h), and an acquisition time of 11 s per step. A graphite secondary monochromator was used to minimize background noise.
Electron microprobe analyses of six thin sections were performed using a SX5FE field emission electron microprobe (CAMECA, Paris, France) at Ruhr-University Bochum. Natural and synthetic minerals were used as reference materials (see Appendix S1 for details). Element distribution maps were acquired using an acceleration voltage of 15 kV, a probe current of about 75 nA, and a fully focused beam. The Xray intensities of Mg Ka, Sr La, Si Ka, Fe Ka and Ca Ka were measured using wavelengthdispersive spectrometers, whereas additional elements (Al, Mn) were recorded using energydispersive X-ray spectrometry with a silicon drift detector. The images were acquired in continuous stage scan mode. They have a resolution of 1367 9 544 to 1863 9 1040 pixels, with a dwell time of 40 ms per pixel.
Microthermometric analyses (n = 22) were carried out using a Linkam THMS600 heatingfreezing stage (Linkam Scientific Instruments Limited, Tadworth, UK) at the Karlsruhe Institute for Technology (KIT) for internal standardization of the fluid inclusions. For the preparation of the fluid petrography, double-polished thick sections (100 to 150 μm) were prepared. The petrographic relation of each fluid inclusion assemblage was carried out by optical and cathodoluminescence microscopy following Goldstein & Reynolds (1994). The documented fluid inclusion assemblages (FIA) were classified as primary inclusions, which are situated on growth zones (p), pseudo-secondary (ps), secondary (s) and isolated inclusions with no genetic relationship (iso; Walter et al., 2015). For each analysis, three repeated heating and freezing cycles were performed to document the final dissolution temperature of ice (T m,ice ) and hydrohalite (T m,hh ), and the homogenization temperature (T h ). For further genetic interpretations, only fluid inclusions which had a triplet measurement between <0.1°C for T m,ice and T m,hh and <1°C for T h were considered. All inclusions and the corresponding data were evaluated for evidence of any post-entrapment modification.

Geochemical analyses
A total of 248 subsample powders were extracted using a hand-held drill. Those with a mixture of calcite and dolomite [based on X-ray diffraction (XRD) and inductively coupled plasmaoptical emission spectrometry (ICP-OES) data] were treated with 0.27 m Di-Na-EDTA to dissolve the calcitic phase .
Carbon-isotope and oxygen-isotope values (d 13 C and d 18 O) were measured by reacting 0.1 AE 0.01 mg of sample powder with phosphoric acid at 70°C. Sample aliquots were measured on a Thermo Finnigan MAT 253 mass spectrometer equipped with a Gasbench II (Thermo Fisher Scientific, Waltham, MA, USA) and a GC PAL autosampler (CTC Analytics AG, Zwingen, Switzerland) at the Ruhr-University Bochum. The internal dolomite and calcite standards were calibrated against IAEA-603, CO-1, CO-8 and NBS 18 with a 1r-reproducibility (SD) of 0.04& and 0.07&, respectively, for d 13 C; and 0.11& and 0.13&, respectively, for d 18 O. All values are reported in & Vienna Pee Dee Formation belemnite (VPDB), and errors as 1r SD.
Strontium isotope ratios ( 87 Sr/ 86 Sr) were analyzed for 53 samples using the method described in Mueller et al. (2020). Ratios were determined using a Thermal Ionization Mass Spectrometer (TIMS) TI-Box (formerly MAT 262; Spectromat, Bremen, Germany) at the Ruhr-University Bochum. Long-term reproducibility was determined using the reference materials NIST NBS 987 and USGS EN-1, and resulted in a 87 Sr/ 86 Sr ratio of 0.710242 AE 0.000001 2r SE and 0.000031 2r SD (n = 505) for NIST NBS 987, and a 87 Sr/ 86 Sr ratio of 0.709159 AE 0.000002 2r SE and 0.000032 2r SD (n = 418) for USGS EN-1. Errors are reported in 2r SE.

Petrography and succession of phases and cements
Input data for the phase sequence in the contact aureole include: (i) field observations (  Table 3). Here, the most essential features in the successions are presented. Refer to Appendix S1 for more details. Dolomite terminology is after Sibley & Gregg (1987).

Fo Fo
Increasing concentration The metamorphic carbonate and silicate succession is overgrown by centimetre-thick bands of undifferentiated phyllosilicates and cut by millimetre-thick up to centimetre-thick low magnesium calcite [low-Mg calcite 1 (LMC 1)] veins with orange luminescence. The succession ends with a third dolomite phase, Dol marble 3 (Figs 4E, 4F and 6D), which has a red luminescence similar to Dol marble 2 and is characterized by elongated crystal morphology. Submillimetresized to millimetre-sized low-Mg calcite 2 (LMC  2) veins cut LMC 1 veins and older metamorphic phases ( Fig. 4C and D). The full metamorphic succession is overgrown by hydrothermal dolomite (Dol 1), meteoric dedolomite (Dedol 1) and aragonite needle cements (AC 1).

Latemar slope and interior
The general cement-paragenetic succession of Latemar slope and interior begins with the deposition of calcitic and aragonitic precursor sediments that underwent early diagenetic, fabricpreserving dolomitization (Contrin Formation). Contrin dolostone consists of a mediumcrystalline planar-s dolomite matrix (Cdol 1; Fig. 4A and B) with abundant recrystallized and poorly preserved Dasycladacean algae casts. Pore space is often occluded by mediumcrystalline to coarse-crystalline, dull red luminescent, zoned Cdol2 cement. The overlying Latemar Sciliar Formation precursor limestones vary depending on the sub-facies (LPF to UTF in Fig. 2) and are generally characterized by a fine-crystalline to medium-crystalline, dominantly microbial calcite matrix (Harris, 1993;Blendinger, 1994;Emmerich et al., 2005) including marine dog-tooth and radiaxial fibrous calcite (RFC) cement, which occludes pore space ( Fig. 7G and H). Based on the data presented here, hydrothermal dolostones consist of four dolomite cement phases. Dol 1 consists of inclusion-rich finecrystalline (dominant at Latemar slope) to poorly zoned inclusion-rich coarse-crystalline (prevalent at Latemar interior) planar-s dolomite ( Fig. 7A to D; Fig. S5C and D). Dol 1 and 2 are separated by a quartz phase (Qz 1), typically forming hypidiomorphic medium-crystalline to coarse-crystalline aggregates. Dol 2 is zoned, fine-crystalline to coarse-crystalline planar-e to nonplanar dolomite ( Fig. 7A and B; Fig. S4A to G). Dol 3 is the volumetrically most significant, fine-crystalline to coarse-crystalline planar-s to nonplanar dolomite and was subdivided into inclusion-rich inner and inclusion-lean outer zones ( Fig. 7E to H). Dol 4 is composed of coarse-crystalline nonplanar saddle dolomite ( Fig. S5E and F

Fluid inclusion thermometry
Detailed petrography by transmitted light microscopy and microthermometry of fluid inclusion assemblages (FIAs) enables classification according to relative age (Goldstein & Reynolds, 1994;Walter et al., 2015). Data are shown in Fig. 8, focusing on primary inclusions hosted in growth zones and pseudo secondary (sensu Walter et al., 2015) inclusions (Table 4). Primary inclusions along crystal growth zones are typically smaller (<5 to 10 μm) than pseudo secondary assemblages (<5 to 20 μm), which are present in (partly) sealed fractures and cross-cut the primary structures of the host minerals.
Carbon-isotope and oxygen-isotope data and radiogenic strontium ratios Geochemical data is shown in Tables 2 and 3, which include: (i) major-element and traceelement concentrations; (ii) d 13 C and d 18 O ranges; and (iii) 87 Sr/ 86 Sr ratios in sample groups. See Appendix S1 for further data (Tables S1 to S4). Oxygen isotope values range

INTERPRETATION AND DISCUSSION
Three data clusters are discussed below: (i) the metamorphic contact aureole, fringing the intrusion (elevation 1000 m a.s.l.); (ii) the Latemar slope, exposing carbonates 4 to 5 km to the north-west of the intrusion (1800 and 2300 m a.s.l.); and (iii) the Latemar interior, 5 to 7 km to the north-west of the intrusion (2400 and 2600 m a.s.l.). First discussed is the petrography from oldest to youngest phases, followed by the geochemistry and finally a brief synopsis.

Petrography and isotope geochemistry
The history of these rocks commences with the deposition of Anisian-Ladinian Contrin and Sciliar precursor sediments in an isolated platform environment, which later (about 240 Myr) record (sub)recent meteoric diagenesis of the exhumed contact aureole and the Latemar carbonates (Fig. 9). The main events that affected the Contrin dolostones and Sciliar limestones are summarized in Figs 9, 10A and 10B.
High-temperature contact metamorphism of early diagenetic dolostones Contact aureole petrography. The Contrin Formation dolostones in the intrusive contact aureole preserve petrographic evidence of a peakretrograde to retrogradeÀmetamorphic cycle: (i) a peak-metamorphic mineral succession; (ii) neomorphosed minerals representing retrograde cooling; (iii) carbonate phases related to tectonic stress; and (iv) subrecent meteoric cements representing exhumation of the carbonates (Fig. 11). The Predazzo subvolcanic body has metamorphosed early diagenetic Contrin dolostones into dolomite marble (Dol marble 1; Fig. 10A), which forms lenses with diameters of tens of metres (Figs 3F and 12A). The dominantly annealed microfabric (sensu Molli et al., 2000) in Dol marble 1 (Fig. 4C) is indicative of a primary recrystallization stage at low to moderate metamorphic temperatures. Apart from locally preserved Dol marble 1, most Contrin marbles were dedolomitized under the influence of structurally-controlled fluids during peakcontact metamorphism (Ferry et al., 2002). The time-integrated fluid flux of about 300 mol fluid/ cm² resulted in transformation of the dolomite marble to periclase-rich (MgO) calcite marbles (Dedol marble 1) at temperatures of 565 to 710°C (Ferry et al., 2002). One possible factor inducing the transition from Dol marble 1 to dedolomite marble is the transition across the dolomite-tocalcite stability window at T = >500°C. At this stage, the Ca/Mg molar ratio is fairly constant, but the system temperature rises (Morrow et al., 1994, and references therein;Huang et al., 2009). This process is also in agreement with the generation of supercritical aqueous fluids with a greatly reduced kinematic viscosity at pressures >22 MPa (MegaPascals). Peakmetamorphic micro-crystalline spinel and forsterite are present as the finely dispersed phase. The Predazzo Intrusion consists of three distinct intrusive pulses, with the oldest pulse composed of a silica saturated shoshonitic melt ( Fig. 2D; Casetta et al., 2018a) and thus probably silica-containing fluids during the first stage of dedolomitization [CaMg(CO 3 ) 2 = MgO + CaCO 3 + CO 2 ]. Spatial variations in peak-metamorphic mineral content were used to determine a vertical direction of fluid flow during peakmetamorphic dedolomitization. With regard to the present mineral content, two dedolomitization reactions are possible: (i) CaMgSi 2 O 6 + CaMg (CO 3 ) 2 = Mg 2 (SiO 4 ) + CaCO 3 (assuming XCO 2 > 0.7); or (ii) CaMg (CO 3 ) 2 = MgO + CaCO 3 (assuming XCO 2 > 0.05).
Brucite  Fig. 9. Simplified temporal evolution of subsequent events affecting the Contrin Formation (not to scale) and Sciliar Formation carbonates in the present study. Temporal uncertainties are denoted by a question mark '?'. For abbreviations see Table 2, Table 3   common at the contact of preserved Dol marble 1 to the brucite-lean Dedol marble 1 zone. This is expressed by submillimetre-thick to decimetre-thick pseudo-bedding foliation of Dol marble 1, Dedol marble 2, Mg-silicates, lizardite and a suite of undifferentiated phyllosilicates (Figs 3G, 4D, 4E, 10A, 12A and 12D). A likely mechanism creating foliation-inducing strain during retrograde metamorphism could be emplacement of the subsequent granitic and shoshonitic intrusive pulses, sensu Casetta et al. (2018a,b). Assuming that this concept holds true, the above petrographic evolution should have occurred over less than 6 Myr (Laurenzi & Vison a, 1996a, 1996bMundil et al., 1996;Visona, 1997).
Neomorphism of inclusion-poor Dol marble 2 is best assigned to the subsequent intrusive pulses and is exemplified by elongated dolomite crystals, which typically occur in the foliated Dedol marble 2 calcite matrix or direct contact to annealed Dol marble 1 fabrics. Dynamic, deformation-driven recrystallization, forming elongated crystals, is well-known from other tectonically active metamorphic settings and is interpreted to represent temperatures of 350 to 400°C in deformed Carrara marbles (Molli et al., 2000;Leiss & Molli, 2003). Petrographic and field evidence suggests that LMC 1 and neomorphosed, inclusion-lean Dol marble 3 mark the second stage of retrogression. These also form elongated crystals but are precipitated in a phase of tectonic activity, which is expressed by their occurrence in veins that pervasively cut older phases (Figs 3G, 4F and 10A). Clinohumite formed during the late stages of retrogression in rims around forsterite crystals.
Rare and isolated fluid inclusions represent the latest stage of retrogression in Dedol marble 2. Fluid inclusion homogenization temperatures from 200 to 220°C may represent the last stage of retrograde recrystallization (Walter et al., 2017), which is also expressed by calcite twinning in Dedol marble 1. Predominantly thintabular to thick-tabular calcite twins are present ( Fig. 4G and Fig. S2A); these may represent a comparable temperature range of about 170 to 200°C (Ferrill et al., 2004). LMC 2 exclusively occurs in submillimetre-sized veins cutting all older phases (Figs 4F and 10A). This stage is followed by dolomite cement (Dol 1). Warped crystal faces in (saddle) Dol 1 indicate a hydrothermal origin ( Fig. 9 and Fig. S2G and H). The moderate impact of Alpine Orogeny in this area of the Dolomite Mountains (Gaetani et al., 1981;Jacquemyn et al., 2014) induced some compressional stress-triggered circulation of hydrothermal fluids, wherein the precipitation of cementation related to hydrothermal fluids ceased. Latest stage meteoric dedolomite and aragonite cements (Dedol 1, AC 1) fill sub-recent fractures and karst-related porosity.
Contact aureole isotope geochemistry. The meta-carbonates in the aureole display a wide range of geochemical signatures. Carbon-isotope, oxygen-isotope and strontium-isotope distributions range between a marine Triassic carbonate and a metamorphic endmember with numerous intermediate stages (Figs 13 and 14). A comparison between Dol marble 1 in the contact aureole with its recrystallized, early diagenetic Contrin dolostone precursor at the base of the Latemar mountain (4 km to the north-west) reveals only moderate differences in d 13 C (1.7 to 2.4& versus 3.1 to 3.3&; Figs 13A and 17A) and d 18 O values (À0.9 to À3.3& versus À2.9 to À3.1&; Figs 13A and 17A). Both phases are outside the range of Anisian carbonate (d 13 C = ca À0.5 to 1.7&), but the oxygen-isotope signature of Dol marble 1 is weakly enriched in 18 O (Anisian d 18 O = À2.0 to À6.5&). The 87 Sr/ 86 Sr ratios of Dol marble 1 yield a wide range (0.707658 to 0.708929) from Triassic marine ratios to much more radiogenic values (Fig. 15). In contrast, Contrin dolostones have Triassic marine  Latemar (Jacquemyn et al., 2014) exclusively reported marine strontium-isotope data. The much wider isotopic range in the same metamorphic phase presented here, merits attention.
Counterintuitively, published bulk carbonisotope and oxygen-isotope data of dolomite and dedolomite marbles suggests that these fall within the range of preserved Triassic marine carbonates (Ferry et al., 2002; see dolomite and calcite data in Fig. 13). Differences in carbonisotope composition between different calcite marbles (with and without brucite) have been assigned to Rayleigh-type processes during dedolomitization (Ferry et al., 2002). The lack of pervasive geochemical re-equilibration during heating towards (high) metamorphic temperatures is of interest.
The observation of what seem only weakly overprinted (marine) isotope values in contact metamorphic marbles allows two basic interpretations.
1 The recrystallization of precursor carbonates occurred in a diffusion-controlled system by fluid films at grain boundaries or along reaction fronts (Pingitore, 1982;Watson & M€ uller, 2009;Jonas et al., 2017;Pederson et al., 2019a). This type of rock-buffered metamorphism has previously been proposed for the Carrara marble and neighbouring dolomitic units, both geochemically indistinguishable from their non-metamorphic precursor lithologies (Cortecci & Lupi, 1994;Cortecci et al., 1999). 2 Recrystallization with larger fluid volumes, but still in a generally rock-buffered system  Korte et al., 2005;Grossman, 2012;Saltzman & Thomas, 2012;Veizer & Prokoph, 2015). The range of seawater associated with early diagenetic Middle/Late Triassic dolostones is indicated in the light green box (from Geske et al., 2012;Chatalov, 2018;Rieder et al., 2019). Data from previous studies (intrusion : Ferry et al., 2002;Latemar: Jacquemyn et al., 2014) is marked in dark green (dolomite) and blue (calcite) boxes. The annotated trends indicate a shift in d 13 C of about À2& from Dol marble 1 towards Dedol marble 2. (B) Data from (A) is set in the context of available pore space, mineralogical processes, metamorphic temperatures and fluid flow. (Greenwood, 1975). Baumgartner & Ferry (1991) suggested that if the time-integrated fluid flux is not extensive, reactants and products coexist along a significant portion of the flow path. This scenario was also insighted (Gallien et al., 2007) explaining seemingly marine carbon-isotope and oxygen-isotope values in dolostones of the Monzoni Intrusive Complex aureole (about 10 km north-east of the Latemar platform). This preservation model depends on a suite of factors often overlooked due to their complex nature: protolith composition, effects of volatilization, temperature of exchange, exchange kinetics, fluid isotope composition, fluid pressure and fluid flux/mixture (Baumgartner & Valley, 2001, and references therein;Duan et al., 2021).
A closer look at the subsequent peakmetamorphic and retrograde phases reveals that Dedol marble 1, Dedol marble 2 and Dol marble 2 (in contrast to Dol marble 1) largely plot in the range of Anisian marine isotope and Ladinian contact metamorphic overprint values (Fig. 13A). A À3& shift of carbon isotope values starting with Dol marble 1, with the most 13 Cdepleted values in Dedol marble 2 and Dol marble 2. This observation contrasts previous work (Ferry et al., 2002) proposing that differences in  Sheppard & Schwarcz (1970) and Friedman & O'Neil (1977) for dolomite and calcite, respectively. The distinct formation temperature ranges (50°C, 100°C, 200°C, 300°C, 500°C and 650°C) are exemplary for the metamorphic temperature range at the Predazzo Intrusion contact aureole obtained from Ferry et al. (2002) and potential late retrograde diagenesis. Note that the oxygen isotope fractionation between both dolomite and calcite, and the fluid is reduced at elevated temperature, thus at 600°C, the d 18 O carbonate-H2O value is only about +1&. The oxygen isotope composition of basaltic fluids and primary magmatic H 2 O is derived from Harmon & Hoefs (1995), Giggenbach (1992) and Vho et al. (2019). The isotopic composition of contact aureole marbles is indicated below the x-axis in green (Dol marbles) and blue (Dedol marbles).  Fig. 15. Oxygen-isotope composition versus strontium-isotope ratios of dolomite and calcite phases compared to Middle/Late Triassic unaltered marine calcites from Korte et al. (2003) and ranges of former studies by Wilson et al. (1990), Blendinger (1997), Blendinger et al. (2011) andJacquemyn et al. (2014). Except for some calcites and less radiogenic dolomites (0.7074), all Latemar samples fit in the Middle/Late Triassic range and agree with former studies. Some marbles from the contact aureole display Triassic values, while the majority are more radiogenic (up to 0.7100). This trend is counterintuitive as the intrusion displays values from 0.7036 to 0.7064 (shoshonitic units, Sloman, 1989;Marrocchino et al., 2009;Jacquemyn et al., 2014;Casetta et al., 2018b) and 0.6000 to 0.6200 (Jacquemyn et al., 2014). No marbles were analyzed at the Latemar slope/interior. the bulk-carbon composition are related to differential volumes of brucite in a given sample. Note, Dedol marble 2 does not contain brucite. In the view of the authors, this prograding trend has not yet been documented. A driving mechanism for this subsequent trend could be the contribution of magmatic carbon from mantlederived CO 2 , which has d 13 C CO2 values between À5 and À8& (Kyser, 1986;Hoefs, 1997). When the isotopic fractionation of 2& between carbon in magma and gaseous CO 2 at igneous temperatures is considered, a d 13 C CO2 range from À3 to À5& for CO 2 from a magmatic origin results (Dietzel & Kirchhoff, 2002). Metamorphic phases at Predazzo Intrusion contact aureole were influenced by magmatic CO 2 , with larger isotopic shifts dominantly inhibited by a large amount of (rock buffered) carbon from the carbonates.
Oxygen-isotope data indicate a shift of about À6& from Dol marble 1 to Dedol marble 1 and a reverse shift to 18 O-enriched values of ca +4& from Dedol marble 1 to Dedol marble 2 (Fig. 13A). Figure 14 shows the d 18 O values of the formation fluid calculated from the d 18 O VPDB values of marbles and the temperature (obtained from Ferry et al., 2002) using the equations of Sheppard & Schwarcz (1970) and Friedman & O'Neil (1977) for dolomite and calcite, respectively, which considers isotopic equilibrium (Vho et al., 2019). The oxygen isotope fractionation between the mineral (dolomite or calcite) and the fluid decreases at elevated temperature, resulting in an d 18 O carbonate-H2O value of 300°C at about +6& and 600°C at about +1&. Thus, at T ≥ 600°C the calculated fluid composition approaches the d 18 O values of the marble (Figs 11 and 14;Ferry et al., 2002). The calculated d 18 O SMOW of the fluid ranges from about À5& at 50°C up to about +30& at 650°C (Fig. 14). A shift to O 18 -enriched fluids may be explained by the admixture of hydrothermal fluid from altered silicates from the igneous intrusion and/or a so-called primary magmatic water (juvenile). However, corresponding d 18 O SMOW values are documented to be a maximum of +10& (e.g. Giggenbach, 1992;Hoefs, 2009). Thus, the extremely O 18 -enriched calculated fluid values of about +30& (SMOW) cannot be explained solely by the leaching/ exchange of oxygen of silicate minerals from the intrusions. During metamorphosis, the oxygen isotope composition of marine Contrin dolostone (d 18 O VPDB = À2.0& AE 2) is preserved (for example, d 18 O VPDB at 650°C: Dol marble 1) due to very low volumes of meteoric or magmatic fluids involved and the volatile character of H 2 O at such high temperatures. Accordingly, the d 18 O distribution of high-temperature marble indicates closed-system conditions. The weak magmatic influence coincides with the observed d 13 C shift in the marbles and the rock-buffered system for carbon, as discussed above.
Dedolomitization during differential fluid flow may also control the carbonate strontium-isotope ratios. Strontium ratios plot between Triassic marine values in Dol marble 1 and Dedol marble 1 (0.707658 to 0.707942), to more common radiogenic values in all subsequent phases (up to 0.710011 in LMC 1). Although no clear phase-specific trend was observed, the most radiogenic ratios occur in carbonate phases rich in retrograde serpentinized silicates (lizardite, forsterite, clinohumite). If this holds true, the strontium source must be highly radiogenic as silicates are typically depleted in strontium compared to carbonates ( Fig. 5D and H). In the absence of additional high-resolution data, it seems unclear how to accommodate the radiogenic strontium and its origin in the metamorphic sequence. Ferry et al. (2002) suggested an H 2 O-rich fluid from a portion of the intrusion cross-cutting the carbonate rocks at depth. Several studies reported strontium-isotope ratios for the different magmatic units of the Predazzo Intrusion below 0.706500 (Fig. 15, arrays 4 to 7). Other workers suggest albitization of basement granite by seawater formation fluids as a process to contribute radiogenic strontium to hydrothermal fluids (Spencer, 1987). For the time being, it seems at least possible that deep-seated basement fluids may have contributed radiogenic strontium during metamorphism. Meteoric dedolomite and aragonite cements, typically occluding late-stage fractures or karst-related vuggy porosity, share similar carbon-isotope values with late marble phases, but are significantly depleted in their oxygen-isotope composition.

Synopsis:
Contact aureole. The peakmetamorphic and retrograde-metamorphic carbonate phases in the Predazzo aureole reveal: (i) complex petrographic and geochemical patterns driven by dedolomitization; (ii) carbonate neomorphism due to interactions with reactive fluids; and (iii) variable fluid-rock ratios. In contrast to a full metamorphic cycle, geochemical proxy data (d 18 O, d 13 C) of these rocks show uncommonly conservative (rock-buffered) behaviour and plot within the range of Middle Triassic marine-isotope values. Strontium-isotope ratios document fluid evolution during neomorphism of Mg-silicates and suggests mixing of fluids from different sources. Processes and products documented here may act as templates for metamorphic-related overprint patterns of carbonate archives commonly occurring over longer timescales and larger spatial extents.
Metamorphism, early-marine diagenetic dolostones and hydrothermal dolomitization at the Latemar slope Latemar slope petrography. Despite the nearideal conditions of the Latemar natural laboratory, the following limitations exist in establishing spatial patterns in cement paragenesis across a carbonate edifice of several kilometres in dimension: (i) inherited facies variability from the basal Contrin Formation (near the intrusion) to the uppermost dolomitized units of the Sciliar Formation (distance of 6 to 7 km; Middle Tepee Facies; Preto et al., 2019); and (ii) differences in outcrop geometry ranging from dolostone bodies (tens to a few hundreds of metres in size) oriented near-vertical relative to slope bedding (Fig. 3D), to smaller (metres to tens of metres) often facies-bound and bedding-bound dolostone units in the Latemar interior ( Fig. 3B and C; Jacquemyn et al., 2015). The dimensions of dolostone bodies generally decrease with increasing distance from the intrusion.
The Contrin Formation in the Dolomite Mountains was previously described as fabricpreserving, early marine-diagenetic dolomite (Gaetani et al., 1981;Sudiro, 2002). Ghost structures of Dasycladacean algae and the dolomicritic nature of these rocks are considered evidence for syndepositional reflux dolomitization ( Fig. 16A and Fig. S4A and B). These dolomites (or very high-Mg calcites) were probably calcian in nature and experienced further stabilization (Nordeng & Sibley, 1994).
At the base of the Latemar isolated platform, about 4 km north-west of the intrusion, laterstage dolomite recrystallization has obscured sedimentary features in the Contrin dolostone (Figs 4A, 4B and 16B) (Jacquemyn, 2013). The fabric is formed by zoned, euhedral, replacement dolomite (Fig. 4B). Fluid inclusion homogenization temperatures (185 to 200°C) suggest hydrothermal recrystallization of the early diagenetic Contrin dolostones (Fig. 8). During the deposition of the Sciliar precursor limestones, Contrin dolostones were affected by synsedimentary extensional tectonics induced by the overburden of the growing platform . Domal uplift during emplacement of the intrusion caused north-west-south-south-east trending fractures which served as zones of weakness for subsequent dyke intrusion (Doglioni, 1983;Jacquemyn et al., 2014). During the cooling of dyke material, thermal contraction opened hydrothermal fluid pathways at the sediment-dyke interface.
The Sciliar Formation limestone precursor facies upsection of the Contrin Formation (Lower Platform Facies, 4 km north-west from Predazzo Intrusion) contains marine cements (radiaxial fibrous calcite, dog-tooth cement) and preserves a series of primary depositional features. With increasing distance from the intrusion, a series of features is observed that is best explained in context with decreasing fluid temperature and fluxes. Jacquemyn et al. (2014) described contact metamorphic to diagenetic, recrystallized metrewide aureoles, in the vicinity of large dykes in outcrops of the Latemar slope (4 to 5 km from the aureole). With increasing distance to dykes, the dolomite cement paragenesis at the Latemar isolated platform is characterized by decreasing Fe-content from early-stage to late-stage dolomite (Jacquemyn et al., 2014). Similarly, Blomme et al. (2017) documented that bulk Fe-content decreases within centimetres to a few metres from the dykes (Figs 9 and 10B). These authors hypothesized that seawater interaction with mafic minerals in dykes caused Fe-enrichment in the dolomitizing fluid.
In all thin sections, Fe-rich Dol 1 typically consists of patchy-luminescent, inclusion-rich, often corroded cores overgrown by zoned or non-zoned rims. Zonation indicates changes in fluid properties during the first phase of dolomitization. Similarly, the precipitation of Qz 1 in dyke-parallel veins and pore space between Dol 1 and Dol 2 suggests transient silica oversaturation at a specific stage between Dol 1 and Dol 2 precipitation. Qz 1 is only observed in the diagenetic halo of dykes, suggesting short-lived fluid pulses from the dykes. Zoned, in part patchyluminescent, Dol 2 occurs as pore-filling cement. Dol 2 has curved crystal faces with sweeping extinction, typical of saddle dolomite (Radke & Mathis, 1980;Wallace & Hood, 2018) forming at fluid temperatures of about 50 to 320°C (Davies & Smith, 2006;Liu et al., 2014). The temperature range (260 to 290°C) of fluid inclusions in Dol 2 (4.5 km from the intrusion) is typically assigned to the deep burial-grade to the lowgrade metamorphic realm. The volumetrically most significant Dol 3 phase consists of patchy-luminescent, non-corroded to partly corroded cores fringed by translucent rims, similar to Dol 1. The most prominent difference between these two phases is the higher Fe-content in Dol 1 (see Appendix S1). Where present as a pore-filling phase in limestones, the crystals show weakly-warped crystal faces with sweeping extinction similar to the Dol 2 saddle dolomites.
The dolomitizing sequence ends with Dol 4, a porosity-occluding saddle dolomite cement, characterized by a few millimetres to one centimetre-sized crystals with warped faces, sweeping extinction and low Fe-concentrations (Jacquemyn et al., 2014). Dol 4 is volumetrically most significant in pores of boxwork dolostone. Boxwork/cell dolostone commonly occurs upsection of the contact between Contrin Formation dolostones and Lower Platform Facies dolostones at about 4 km from the intrusion. Zebra dolomite accompanied by dilational fracturing is present (Jacquemyn et al., 2014). Boxwork dolomite does occur bed parallel or, in many cases, non-strata bound and rather chaotic (Fig. 3E). Boxwork dolomite may form as the result of previously fractured precursor limestones affected by pressure release during dolomitization and later dissolution of limestone clasts (Figs 3E, 12B and 12E;Dunsmore, 1973;Wallace & Hood, 2018). An alternative formation mode is dissolution in permeable limestone or dolostone beds (Morrow, 2014).
After a hiatus in the cement stratigraphy, fluids with lower Mg/Ca ratios, or an ephemeral temperature rise at fairly constant Mg/Ca ratios, led to precipitation of millimetre-sized to several centimetre-sized blocky LMC 1 in all dolomitized limestone samples. Fluid inclusion data from the Latemar slope (4.5 km from the intrusion) document elevated fluid temperatures (188 to 195°C). Blocky and fibrous LMC 2 filling fissures in dykes likely also precipitated during this stage or shortly thereafter. Some growth zones in several dolomite crystals show dedolomitization features ( Fig. 7C and D). The resulting pore space is typically filled by Feoxides, which increase the bulk Fe-content of dolomite phases and limit the applicability of trace element data. The latest stage LMC 3 and HMC 1 cements are rare and represent a moderate petrographic meteoric/vadose overprint (Figs 9 and 10B).
Latemar slope isotope geochemistry. The ongoing controversy regarding the in part 13 Cenriched marine signatures of the Latemar Sciliar limestones [see Jacquemyn (2013) and Jacquemyn et al. (2014) for a detailed discussion] is acknowledged. The argument is that Latemar limestones with 13 C-enriched signatures, relative to coeval carbonate platforms in the Dolomite Mountains, might record early stages of hydrothermal CO 2 exhalation at the seafloor pre-dating the intrusion of igneous bodies. Based on bulk, and more recently, cementspecific analyses, Triassic marine values have been reported for carbon, oxygen and strontium isotopes (Wilson, 1989;Carmichael, 2006;Jacquemyn et al., 2014). Heavier carbon-isotope values, relative to Carnian seawater, were first reported by Carmichael et al. (2008). Jacquemyn et al. (2014) argued that most samples reflect the Carnian seawater isotope composition during Sciliar carbonate deposition. Phase-specific data presented here further support this interpretation. An alternative source of heavy 13 C might be syndepositional methanogenesis of organic material in lagoon shoal-settings (Teal et al., 2000).
Dolomitized Latemar slope oolithic ironstones yield the most positive d 13 C values (up to +5&; Fig. 17A), as do Dol 1 (Dol 1 max . = +4.9&; Fig. 17A) samples in the slope, 4 to 5 km from the intrusion. All of these samples were collected in close vicinity (i.e. within decimetres to metres) of dykes. This study observed a difference of +2.5& between the limestone precursor facies and Dol 1, and a difference of +1& between the limestone precursor and Dol 2. Later dolomite phases Dol 3 and Saddle Dol 4 lack this pattern, independent of the sampling location.
Werfen Formation marlstone bulk samples collected below the Contrin Formation in the direct vicinity of dykes (3.5 to 4.0 km north-west of Predazzo Intrusion) yield diagenetically altered d 18 O values (À11.0 to À12.5&). Generally, no significant burialÀdiagenetic or meteoricÀ diagenetic overprint of the d 13 C dolomite and d 18 O dolomite signatures was observed (Fig. 17A). Exceptions include a few samples (n = 5) of Dol 1 and later stage dolomites (Dol 3 and Dol 4). Diagenetic overprint of Sciliar d 18 O calcite signatures is reflected in late calcite cements occluding minute pores and vugs. This is best visible in LMC 1 cements with d 18 O values as low as À12.3&. Nearly all sampled phases, including early diagenetic Contrin dolomites from the base of Latemar slope, plot within the range of Triassic marine precipitates.
The concept of geochemical patterns in Dol 1 and Dol 2 precipitated from intrusion-derived fluids merits attention. Dolomitizing fluids may migrate at basin-scale, with fluid flow driven by a range of processes (Braithwaite et al., 2004). Still, the setting described here, i.e. metamorphic dedolomitization acting as a source of Mg in dolomitizing fluids, arguably represents a less-commonly described process. It seems likely that some of the Mg, released during peak-metamorphic dedolomitization, migrated with the dolomitizing fluid. This interpretation is strengthened by the observation that only some Mg at the contact aureole was sequestered within the neomorphosed Mg-silicates.
The strontium-isotope ratio of most Sciliar Formation phases, independent of their spatial position, plot in the range of biogenic and abiogenic marine calcite (Korte et al., 2003;Fig. 15). The data reveal no difference between most precursor limestones and dolomite phases sampled in the Latemar slope. Generally, only two Latemar slope dolomite samples are less radiogenic compared to the global seawater curve. Similar patterns have been observed in the context of shoshonitic or granitic units of the Predazzo Intrusion (Jacquemyn et al., 2014). Similarly, only three samples are more radiogenic than predicted Triassic marine calcites. The source of the more radiogenic strontium remains unclear at present. For the time being, the question of whether there is a relation between more radiogenic strontium ratios in samples collected in the contact aureole and samples from the Latemar isolated platform remains unresolved. If the radiogenic strontium was supplied by a deep-seated basement fluid rising through intrusion-related pathways, it would seem likely that both the contact aureole and the dolomitized strata of the Sciliar Formation at Latemar should be affected.
Synopsis: Latemar slope. Metasomatizing fluids trigger the dolomitization of Sciliar Formation limestones now exposed at the slope of the Latemar (about 4 km from the intrusion). The volume and size of hydrothermal dolostone rock bodies embedded in limestones decrease systematically (from hundreds to a few metres) with increasing distance (kilometres) from the intrusion, arguably reflecting decreasing fluid-rock interaction. At the metre-scale, the dolomite cement paragenesis is characterized by decreasing Fe-content from early to late phases with increasing distance to dykes. At the submillimetre scale, fluid inclusion homogenization temperatures indicate a hydrothermal temperature regime between 180 and 290°C. The majority of paragenetic phases yield oxygenisotope and strontium-isotope compositions typical of Triassic marine carbonates. This is remarkable, particularly considering that the carbonate oxygen-isotope system is prone to temperature-controlled fractionation. However, temperature-related fluidÀcarbonate fractionation of oxygen decreases at higher temperatures (Sheppard & Scharcz, 1970). A difference of d 13 C values of +2.5& between limestone precursor and Dol 1, and +1& between limestone precursor and Dol 2 phase was observed. Dolomite phases Dol 3 and Saddle Dol 4 taken from the same specimen lack this trend.
Relation between contact metamorphism and dolomitization in the Latemar isolated platform interior Latemar interior petrography. The isolated platform interior is spatially situated at 5 to 7 km lateral distance and 1.4 to 1.6 km in altitude from the intrusion and represents the low-temperature end member of the transect. The Sciliar Formation limestone precursor facies typically lacks the metre-wide aureoles fringing large dykes at sites closer to the intrusion. Sciliar limestones show limited-to-no evidence for petrographic alteration at the contact to intrusive dykes due to the comparably small volumes of low-temperature hydrothermal fluids involved. Alternative interpretations include higher crystallization rates due to a lower temperature gradient of the melt further from the source.
Replacive dolostones in Latemar's interior lack boxwork dolostone fabrics with saddle dolomite and later LMC 1. In the platform interior, limestone-dolostone transition zones are represented by vein swarms filled by euhedral dolomite, which cross-cut the precursor limestone in zones several decimetres wide (Figs 3C, 12C and 12F). Fluid-inclusion homogenization temperatures of 54 to 72°C in Dol 2 (at 6.5 km from the intrusion) are markedly cooler compared to those in Dol 2 (260 to 290°C), 2.5 km closer to the intrusion. This may indicate a time-equivalent temperature evolution with increasing distance from the fluid source for dolomitization. Interestingly, the range of 54 to 72°C is also in agreement with carbonate-clumped isotope data (Ferry et al., 2011) from the Latemar interior. Further research to verify fluid inclusion data confidence and to detect a spatial temperature trend at Latemar is underway.
The directly comparable cement stratigraphy (Dol 1, 2 and 3) in outcrops at the Latemar slope and the platform interior supports the interpretation of contemporaneous dolomitization events. Differences are mainly related to: (i) the spatial distance between the sample and an intrusive dyke; (ii) Fe-content; and potentially (iii) decreasing fluid inclusion homogenization temperatures towards the Latemar interior.
Latemar interior isotope geochemistry. Metamorphic dedolomitization at the intrusive contact acted as a Mg source for dolomitizing fluids in the Latemar interior (Jacquemyn et al., 2014(Jacquemyn et al., , 2015. A spatially resolved carbon-isotope gradient is recognized when comparing outcrops of Latemar slope dolostones with those of the more distal Latemar interior. In agreement with petrographic evidence for decreasing fluid temperature, the metamorphic influence on d 13 C dolomite and d 18 O dolomite values become increasingly less important, with a difference of about +1& between all limestone and Dol 1 samples (Fig. 17B). All subsequent dolomite phases (Dol 2, Dol 3 and Dol 4) lack this pattern, suggesting that d 13 C values are rock-buffered. The weakly 13 C-enriched data from the Latemar interior are likely related to increasing distance to the metasomatized fluid source. Remarkably, three subsamples of a partly recrystallized Sciliar limestone, from the direct vicinity (<1 cm) of a dyke (5 km from the intrusion), have a similarly positive d 13 C calcite signature (3.8 to 3.9&; Fig. 17B). All of these patterns point to the complexity of the system, even when samples are placed in a temporal and spatial context. Jacquemyn et al. (2014) argued that dedolomitization of the aureole dolomites alone is likely insufficient to provide the amount of Mg required to dolomitize significant volumes of limestone. These authors instead suggested that the characteristic Carnian seawater signatures in hydrothermal bulk replacement dolomite indicate a seawater contribution. If this holds true, two independent Mg sources best explain the early diagenetic dolomitization of the Latemar: (i) Mg from the contact metamorphic dedolomitization; and (ii) Mg from Carnian seawater.
When comparing the oxygen-isotope values of limestone precursors and dolomite phases, the most relevant factor is temperature-controlled fractionation (up to about 100°C), which adds to mineralogical fractionation between calcite and dolomite (up to 3&; Swart, 2015). Care must be taken, however, as mineral-dependent fractionation is most significant at ambient seawater temperatures and decreases at higher fluid temperatures. Mineral-dependent fractionation is best documented in replacement dolomite Dol 3 exposed in the Latemar interior outcrops (Fig. 17B). Similar to samples from the Latemar slope, the d 18 O dolomite values (n = 9) of Dol 3 and Saddle Dol 4, are 18 O-depleted compared to their limestone precursors (Fig. 17B).
Similar to Latemar slope samples, the strontium-isotope ratios of most Sciliar Formation phases, independent of their spatial position, plot in the range of marine calcite (Korte et al., 2003;Fig. 15). Only a limited number (n = 2) of late diagenetic LMC 2 samples are less radiogenic than predicted by the global seawater curve, implying that diagenesis and moderate hydrothermal dolomitization did not obscure the marine archive. Despite the fact that the Latemar slope and interior represent two distinct stratigraphic sections, it seems unlikely (but not impossible) that these trends in carbonate isotope values reflect spatial differences in seawater geochemistry.
Synopsis: Latemar interior. The Latemar platform interior, located 5 to 7 km from the intrusion, lacks metre-wide recrystallized contact aureole fringing dykes. Dolomitizing fluid temperatures range from 50 to 70°C. A spatial carbon isotopic gradient between the Latemar slope and interior dolostones is observed. The influence of intrusion-related hydrothermal fluids in the d 13 C dolomite signatures of the Latemar interior is suppressed (+1& between limestone and Dol 1). The oxygen-isotope values of Sciliar limestones and hydrothermal dolomites are controlled by mineralogical and temperaturedependent fractionation, relevant in a temperature range from 40 to 80°C.

Significance for deep-time carbonate archive research
The data and interpretations documented here are relevant for carbonate archive research because of several reasons. The Phanerozoic Eonothem offers sufficient well-preserved biogenic and abiogenic carbonate archives and, hence, samples with evidence for secondary alteration are discarded. Nevertheless, Swart (2015) and others argued that many of the proxy records from ancient carbonate archives may not represent the depositional environment, but are rather (early) diagenetic signals. Even in some sub-recent foraminifera tests and other carbonate archives, subtle yet significant diagenetic effects were found (Perrin & Smith, 2007;Bernard et al., 2017;Lange et al., 2018;Pederson et al., 2019b). Although there is no correlation between the geological age of a given carbonate archive and its degree of preservation, these issues become increasingly significant when dealing with the 'deep-time' rock record. Specifically, meta-carbonate rocks that have undergone pervasive recrystallization at greater depth and under fluid temperatures of >300°C represent the frontier in this regard. In many natural systems, advective or diffusive transport creates fluid-buffered conditions. A common perspective is that 'wet´metamorphism eradicates any useful environmental signal. The issue is particularly relevant for the low-grade to high-grade Proterozoic and Archean carbonate archives. Examples include Archean stromatolitic carbonates in greenstone belts (Condie, 1981;Klein, 2005) and the 3.8 Ga Isua metacarbonates, heated to temperatures of 460 to 600°C, that contain debated evidence for the earliest life on Earth (van Zuilen et al., 2002;Nutman et al., 2016). Archean metacarbonate archives have often undergone multiple episodes of metamorphism with barely distinguishable episodes of regional and contact metamorphism (Hayashi et al., 2000;Nutman et al., 2010;Sizova et al., 2018).
The present study in the Latemar offers important insights into metamorphic to diagenetic fluid-rock interaction gradients between some tens of degrees Celsius and 700°C in a lowtemperature to contact metamorphic hydrothermal regime. Both strongly recrystallized marbles in the contact aureole and limestones affected by warm hydrothermal fluids in the platform interior display d 13 C and d 18 O values similar to (or overlapping with) Triassic seawater. Textbook interpretations in the sense of 'petrographic alteration of a given carbonate is accompanied by its geochemical reequilibration' are not applicable here. Following arguments brought forward by others (Ferry et al., 2002;Jacquemyn et al., 2014), this study finds evidence for decoupled (in both time and space) petrographic and geochemical alteration fronts in the Latemar rocks. It is emphasized, however, that they are fully aware of the differences in terms of pressure regimes and heat transfer rates between contact and regional metamorphism. Thus, care must be taken to avoid overly simplistic interpretations based on the data shown here (see also discussion in Baumgartner & Valley, 2001;Ferry et al., 2011Ferry et al., , 2014. Acknowledging all of these issues, it is argued that the alteration gradients at the Latemar offer fundamental insights into processes and products that have relevance for deep-time carbonate archive research.

CONCLUSIONS
This study documents a complex suite of processes ranging from contact metamorphic to hydrothermal dolomitization of precursor limestones in the natural laboratory of the Triassic Latemar isolated platform and the nearby subvolcanic Predazzo Intrusion (northern Italy). A transect from the contact metamorphic aureole towards the low-temperature hydrothermal endmember characterizes carbonates based on their petrographic, paragenetic and geochemical criteria. Spatially restricted dolostone bodies are embedded in marine precursor limestones and allow a direct comparison of processes and products.
At the contact metamorphic aureole, three dolomite marbles, two dedolomite marbles, several magnesium hydroxides, silicates and a series of hydrothermal to late-stage meteoric cements are observed. In strong contrast with petrographic and mineralogical evidence for a full metamorphic cycle, carbon-isotope and oxygen-isotope values seem to reflect preserved Triassic 'marine' values. Strontium-isotope ratios suggest multiple fluid sources. Partial dolomitization of carbonates occurred between 4 km and 7 km from the intrusive heat source and was driven by hydrothermal fluids circulating through north-west to southeast trending dykes radiating from the Predazzo Intrusion into the Latemar isolated platform. Four hydrothermal dolomite phases formed. These features are considered evidence for decreasing fluid temperature, fluid flow, and Mg and Fe concentrations with increasing distance from the metamorphic heat source.
Oxygen-isotope values of Latemar Sciliar limestones and dolostones, as well as early diagenetic Contrin dolomites at its base, plot in the range of Triassic marine precipitates. A difference of up to 2.5& in d 13 C values is observed between limestone precursors and early-stage hydrothermal dolomite phases. This shift is less pronounced in the Latemar interior, which has a difference of 1& between limestones and dolostones. The moderately 13 C-enriched values of Latemar interior carbonates are best explained by decreasing fluid temperature with increasing distance to the metasomatizing fluid source. Late-stage dolomite phases overlap with precursor limestones in terms of their isotope geochemistry. This pattern arguably reflects coeval metamorphic dedolomitization releasing Mg into the dolomitizing fluids. The oxygen-isotope composition of Sciliar limestones and later hydrothermal dolomites is predominantly controlled by mineralogical fractionation from calcite to dolomite and temperature-dependent fractionation. The strontium-isotope ratios of most Sciliar Formation phases plot in the biogenic and abiogenic marine calcite range, independent of their spatial position.
This study is relevant as it documents that even significant petrographic alteration of carbonate archives does not per se invoke geochemical re-equilibration. The processes and products documented are best explained through rockbuffered metamorphic and diagenetic alteration. Data provided offer a framework against which existing and subsequent work can be placed to extract quantitative results, patterns and methods for the community.

Supporting Information
Additional information may be found in the online version of this article: Appendix S1. Supporting information on methodology, sampling locations and circumstantial data on major and trace elements, geochemical, petrographical and fluid inclusion results. Fig. S1. Geological map of the Latemar Atoll (modified from Jacquemyn, 2013). Fig. S2. Additional transmitted light images placed against cathodoluminescence properties of different paragenetic phases in partly dedolomitized Contrin Formation dolomite marble from the contact metamorphic aureole of the Predazzo Intrusion in the Canzoccoli quarry. Fig. S3. Additional microprobe element maps of the contact metamorphosed Contrin Formation near the Predazzo Intrusion. Fig. S4. Microprobe element maps of Fe-rich dolomite phases Dol 1 and Dol 2 at Latemar slope. Fig. S5. Additional transmitted light images were placed against cathodoluminescence properties of different paragenetic phases in dolomitized limestone facies of the Latemar atoll. Table S1. Carbon, oxygen and 87 Sr/ 86 Sr isotopic composition as well as major and trace element contents of dolomite, calcite, and aragonite phases for the Contact aureole. Table S2. Carbon, oxygen and 87 Sr/ 86 Sr isotopic composition as well as major and trace element contents of dolomite, calcite, and aragonite phases for the Latemar slope. Table S3. Carbon, oxygen and 87 Sr/ 86 Sr isotopic composition as well as major and trace element contents of dolomite, calcite, and aragonite phases for the Latemar interior. Table S4. Minimum and maximum values for carbon, oxygen, and 87 Sr/ 86 Sr isotopic composition as well as calculated means for Carbon and oxygen of all analyzed phases. Table S5. Fluid inclusion data of all analyzed phases. Table S6. Temperature-dependent fluid composition (VSMOW) based on d 18 O VPDB of Predazzo Intrusion contact aureole marble samples.