From Early Contraction to Post-Folding Fluid Evolution in the Frontal Part of the B ó ixols Thrust Sheet ( Southern Pyrenees ) as Revealed by the Texture and Geochemistry of Calcite Cements

Structural, petrological and geochemical (δ13C, δ18O, clumped isotopes, 87Sr/86Sr and ICP-MS) analyses of fracture-related calcite cements and host rocks are used to establish a fluid-flow evolution model for the frontal part of the Bóixols thrust sheet (Southern Pyrenees). Five fracture events associated with the growth of the thrust-related Bóixols anticline and Coll de Nargó syncline during the Alpine orogeny are distinguished. These fractures were cemented with four generations of calcite cements, revealing that such structures allowed the migration of different marine and meteoric fluids through time. During the early contraction stage, Lower Cretaceous seawater circulated and precipitated calcite cement Cc1, whereas during the main folding stage, the system opened to meteoric waters, which mixed with the connate seawater and precipitated calcite cement Cc2. Afterwards, during the post-folding stages, connate evaporated marine fluids circulated through newly formed NW-SE and NE-SW conjugate fractures and later through strike-slip faults and precipitated calcite cements Cc3 and Cc4. The overall paragenetic sequence reveals the progressive dewatering of Cretaceous marine host sediments during progressive burial, deformation and fold tightening and the input of meteoric waters only during the main folding stage. This study illustrates the changes of fracture systems and the associated fluid-flow regimes during the evolution of fault-associated folds during orogenic growth.


Geological Setting
The Pyrenees formed by the continental collision between the Iberian and European plates and consist of an asymmetrical and doubly-verging orogenic system generated from the Late Cretaceous to the Oligocene [60][61][62][63]. This orogen is constituted by an antiformal stack of basement-involved thrusts (i.e., Axial Zone) surrounded by two fold and thrust belts that were transported to the north and south over the Aquitanian and Ebro foreland basins, respectively [61,64,65]. The South-Central

Geological Setting
The Pyrenees formed by the continental collision between the Iberian and European plates and consist of an asymmetrical and doubly-verging orogenic system generated from the Late Cretaceous to the Oligocene [60][61][62][63]. This orogen is constituted by an antiformal stack of basement-involved thrusts (i.e., Axial Zone) surrounded by two fold and thrust belts that were transported to the north and south over the Aquitanian and Ebro foreland basins, respectively [61,64,65]. The South-Central Pyrenean Unit consists of three south-verging imbricated thrust sheets detached over Triassic evaporites: Bóixols, Montsec and Serres Marginals [66]. The Bóixols thrust sheet represents the northernmost and oldest emplaced structure and was originated by the inversion of the Organyà extensional basin during the Upper Santonian [59,[67][68][69]. It is bounded to the south by the Bóixols thrust and to the north by the Morreres back-thrust and is characterized by a set of structures roughly striking E-W [66]. The study area is located in the frontal part of the Bóixols thrust sheet and comprises the southern limb of the Bóixols anticline and the northern limb of the Coll de Nargó syncline ( Figure 1). Existing stratigraphic data [70,71] reveals that the main stage of development of the Bóixols thrust was from the Campanian (83.6-72.1 Ma) to the late Maastrichtian (72.1-66 Ma) [72]. Additionally, authigenic illite dating of the above-mentioned fault gouge indicates an age of 71.2 ± 6.4 Ma [73]. This result is consistent with the age estimated through stratigraphic data, confirming the onset of the Pyrenean convergence and associated deformation by the Late Cretaceous.

Lithology and Petrography
The stratigraphic record in the Bóixols thrust sheet has been divided into four main sequences related to the main tectonic events ( Figure 2) [67,72,74]: (1) The pre-rift sequence mainly consists of evaporites and clays of the Triassic Keuper facies, which acted as a detachment level of the Bóixols thrust [75], and Jurassic dolostones; (2) The Lower Cretaceous syn-rift sequence is represented by platform carbonates and basinal marls [67,69]; (3) The post-rift sequence, with an age from Upper Cenomanian to Early Santonian, is composed of limestones, sandstones and marls [69]; and, finally, (4) the syn-orogenic sequence is represented by Upper Santonian to Campanian turbiditic deposits [76], Late Campanian to Maastrichtian shallow marine sequences and uppermost Maastrichtian to Paleocene fluvial lacustrine red beds. The succession cropping out in the study area (Figures 3 and 4) is described as follows. The Jurassic pre-rift sequence mainly consists of limestones, marls and dolostones with a thickness ranging from 600 to 1,200 m [72]. However, this unit often appears incomplete in the study area because of erosion [69,72]. These sediments were incorporated into the hanging wall of the Bóixols thrust sheet during the Alpine orogeny. The Lower Cretaceous syn-rift Lluçà Formation (Aptian-Albian in age) consists of a rhythmic alternation of dark grey marls, marly limestones and limestones containing silt to fine quartz grains and sporadic glauconite grains, representing water depth variations. The abundant fossiliferous content (ammonites, sponges, rudists and isolated calcispheres) suggests that these sediments were deposited from a shallow to deeper marine environment below the wave action zone [69].The Upper Cretaceous post-rift Santa Fe Formation deposited prior to the Alpine compression [58]. It is bounded by two unconformities and consists of a rigid 30-m thick succession of grey homogeneous bioclastic wackestones constituted of miliolida, bivalves, bryozoans, echinoderms, Textularia and Praealveolina. Bio-erosion evidence on fossil fragments suggests that deposition of this unit took place during low sedimentation rates and/or shallow conditions. The uppermost Cretaceous to Paleocene syn-orogenic Garumnian facies (i.e., Tremp Formation) are composed of fluvial-alluvial and lacustrine facies comprising red mudstones, limestones, sandstones, conglomerates and versicolored clays ( Figure 5A-G). Conglomerates are polymictic, clast-supported and formed by 70-80% of Mesozoic extrabasinal calcareous pebbles, characterized by a wide grain size range (1-20 cm) and shape varying from sub-angular to sub-rounded with low to high sphericity. The remaining part is constituted by 10-30% matrix, 10-20% calcite cement and negligible porosity (0-5%). Red-brick floodplain mudstones are characterized by high iron oxide concentration. Lacustrine limestones generally occur interbedded with red and nodular mudstones. Petrographically, they consist of wackestones composed of fragments of bivalves and gastropods, which originally were composed of low-Mg calcite [77]. Locally they change to oncolitic packstones ( Figure 5H) displaying a peloidal fabric related to microbial processes. Fine and coarse-grained sandstones are stacked in tabular and channelized bodies. Versicolored clays appear altered by oxidereduction processes related to paleo-soils. Locally, packstone with Microcodium is also observed ( Figure 5I).     The uppermost Cretaceous to Paleocene syn-orogenic Garumnian facies (i.e., Tremp Formation) are composed of fluvial-alluvial and lacustrine facies comprising red mudstones, limestones, sandstones, conglomerates and versicolored clays (Figure 5a-g). Conglomerates are polymictic, clast-supported and formed by 70-80% of Mesozoic extrabasinal calcareous pebbles, characterized by a wide grain size range (1-20 cm) and shape varying from sub-angular to sub-rounded with low to high sphericity. The remaining part is constituted by 10-30% matrix, 10-20% calcite cement and negligible porosity (0-5%). Red-brick floodplain mudstones are characterized by high iron oxide concentration. Lacustrine limestones generally occur interbedded with red and nodular mudstones. Petrographically, they consist of wackestones composed of fragments of bivalves and gastropods, which originally were composed of low-Mg calcite [77]. Locally they change to oncolitic packstones ( Figure 5h) displaying a peloidal fabric related to microbial processes. Fine and coarse-grained sandstones are stacked in tabular and channelized bodies. Versicolored clays appear altered by oxide-reduction processes related to paleo-soils. Locally, packstone with Microcodium is also observed (Figure 5i).

The Coll de Nargó Syncline
The Coll de Nargó syncline is a narrow, asymmetric and E-W oriented fold formed in response to the Bóixols thrusting and located at the north-eastern boundary of the Montsec thrust sheet (Figures 3 and 4). It includes the Areny Formation (Maastrichtian) and the Garumnian facies

The Coll de Nargó Syncline
The Coll de Nargó syncline is a narrow, asymmetric and E-W oriented fold formed in response to the Bóixols thrusting and located at the north-eastern boundary of the Montsec thrust sheet (Figures 3  and 4). It includes the Areny Formation (Maastrichtian) and the Garumnian facies (Paleocene), which unconformably overly the Santa Fe Formation (Cenomanian). In the northern limb, bedding ranges from vertical to slightly overturned, while it gently dips (30-40 • ) towards the north in the southern flank ( Figure 4). The syntectonic deposition of the Garumnian facies is highlighted through successive unconformities registered along the sedimentary record.

Methods
This study has been carried out by an integrated geological approach. Structural data were collected in the field including fracture and bedding orientation, types, crosscutting relationships between different fracture sets and kinematics when possible. Structural data were plotted and analyzed with the software Win-Tensor (v.5.8.8) [78]. Host rocks and vein calcite cements were systematically sampled for petrographical and geochemical analyses.

Petrography
In order to distinguish different calcite cement generations, forty polished thin-sections (notated as CN1 to CN40) were prepared and analyzed with a Zeiss Axiophot optical microscope and a Technosyn Cold Cathodoluminescence microscope, model 8200 Mk5-1 operating between 16-18 kV and 250 µA gun current. The terminology of vein types and vein microstructures is based on the classification of Reference [79].

Carbon and Oxygen Isotopes
The calcite cements and their related carbonate host rocks were carefully sampled for carbon and oxygen isotopic analysis using a 400/500 µm-diameter dental drill to extract 50-100 µg of powder. Calcite powders were reacted with 100% phosphoric acid for two minutes at 70 • C. The resultant CO 2 was analyzed using an automated Kiel Carbonate Device attached to a Thermal Ionization Mass Spectrometer Thermo Electron MAT-252 (Thermo Fisher Scientific, Bremen, Germany) following the method of [80]. The International Standard NBS-18 and the internal standard RC-1, traceable to the International Standard NBS-19, were used for calibration. The standard deviation is ±0.03‰ for δ 13 C and ±0.05‰ for δ 18 O expressed with respect to the VPDB standard (Vienna Pee Dee Belemnite). The list of analyzed samples is shown in Table 1.

Clumped Isotope Thermometry
Clumped isotope thermometry was carried out in order to calculate temperatures and δ 18 O values of the fluids from which the different generations of calcite cements precipitated. This technique is giving good results in this kind of study [16,17,81] and it is becoming a key method in order to establish temperature and fluid composition in carbonate cements, where usually fluid inclusion microthermometric analyses cannot be performed due to stretching. Samples CN33, CN38, CN20 and CN15 were chosen as representative of each calcite cement generation (Cc1 to Cc4, respectively). For this, 2-3 mg aliquots from cements were measured with the Imperial Batch Extraction system (IBEX), an automated line developed at Imperial College. Each sample was dropped in 105% phosphoric acid at 90 • C and reacted for 30 min. The reactant CO 2 was separated using a poropak-Q column, and transferred into the bellows of a Thermo Scientific MAT 253 mass spectrometer (Thermo Fisher GmbH, Bremen, Germany). The characterization of a replicate consisted of 8 acquisitions in dual inlet mode with 7 cycles per acquisition. The post-acquisition processing was completed with a software for clumped isotope analysis named Easotope [82]. ∆ 47 values were corrected for isotope fractionation during phosphoric acid digestion employing a phosphoric acid correction of 0.069 ‰ at 90 • C for calcite [83], the data were also corrected for non-linearity applying the heated gas method [84] and projected into the reference frame of [85]. Carbonate δ 18 O values were calculated with the acid fractionation factors of [86]. Samples were measured two times, except CN38 with only one replicate, and the average result was converted to temperatures with the calibration method of [87].

Strontium Isotopes
Ten representative samples of calcite cements (samples CN6, CN12, CN15, CN18, CN20, CN27, CN33, CN38) and host rocks (CN12 and CN16) were sampled for 87 Sr/ 86 Sr analyses. Powdered samples were completely dissolved in 5 mL of 10% acetic acid, dried and again dissolved in 1 mL of 65% HNO 3 . After evaporation, the resulted solid residue was diluted in 3 mL of 3M HNO 3 to be charged in the chromatographic columns in order to obtain the Sr. In the chromatographic columns, a SrResinTM (Triskem Internacional) (crown-ether (4,4 (5 )-di-t-butylcyclohexano-18-crown-6)) resin was used and Sr was recovered using 0.05M HNO 3 as an eluent. After evaporation, samples were loaded onto a Re filament with 1 µL of 1 M phosphoric acid and 2 µL of Ta 2 O 5 . Isotopic ratio measurements were carried out in a TIMS-Phoenix mass spectrometer (Isotopx, Cheshire, UK) with a dynamic multicollection during 10 blocks of 16 cycles each one, with a 88 Sr beam intensity of 3 V. Isotopic ratios were corrected from 87 Rb interferences and normalized using the measured value of 88 Sr/ 86 Sr = 0.1194 to correct for the possible mass fractionation during loading and analysis of the sample. During sample analysis, the isotopic standard NBS-987 was measured six times, obtaining a media of 0.710243 and a standard deviation 2σ of 0.000009. These values have been used for the correction of the analyzed values in the samples. The analytical error in the 87 Sr/ 86 Sr ratio, referred to two standard deviations, is 0.01%. The standard error or internal precision is 0.000003.

Elemental Composition
Eight samples (CN6, CN12, CN15, CN18, CN20, CN27, CN33, CN38), two from each calcite cement, were analyzed for major, trace and rare earth element geochemistry using HR-ICP-MS in order to obtain a range of compositions for each cement generation. This technique was chosen due to its higher resolution compared to other methods (i.e., electron microprobe) and the possibility to perform a multielemental analysis including rare earth elements (REE), which give important information about fluid origin and redox conditions. Trims were sampled using a 400/500 µm-diameter dental drill to extract 100 mg of powder. Powdered samples were dried at 40 • C during 24 h and later 100 mg of sample were acid digested in closed polytetrafluoroethylene (PTFE) vessels with a combination of HNO 3 + HF + HClO 4 (2.5 mL: 5 mL: 2.5 mL v/v). The samples were evaporated and, 1 mL of HNO 3 was added to make a double evaporation. Finally, the sample was re-dissolved and diluted with MilliQ water (18.2 MΩ cm −1 ) and 1 mL of HNO 3 in a 100 mL volume flask. Analyses were performed using a high resolution inductively coupled plasma-mass spectrometry (HR-ICP-MS, model Element XR, Thermo Fisher Scientific, Bremen, Germany). In order to improve the sensitivity of the ICP-MS, a tuning solution containing 1 g L −1 Li, B, Na, K, Sc, Fe, Co, Cu, Ga, Y, Rh, In, Ba, Tl, U was used, and as internal standard, 20 mg L −1 of a monoelemental solution of 115 In. Reference materials are the BCS-CRM n • 393 (ECRM 752-1) limestone, JA-2 Andesite and JB-3 Basalt. The precision of the results was expressed in terms of two standard deviations of a set of eight reference materials measurements (reference material JA-2), whereas accuracy (%) was calculated using the absolute value of the difference between the measured values obtained during the analysis and the certified values of a set of eight reference material analysis (reference material BCS-CRM n • 393 for major oxides and JA-2 for trace elements). The detection limit (DL) was calculated as three times the standard deviation of the average of ten blanks. Detection limits and two standard deviations of each element are shown in Supplementary Data. Several commercial solutions were used in order to perform the different calibration curves. Calibration standards for Fe, Al, P, Ti, Na, K, Ca, Mg were Monoelemental Solutions (10,000 µg/mL) of Inorganic Ventures; for Sr, Ba, Li, Be, V, Cr, Ni, Cu, Ga, As, Se, Rb, Cs

Fracture Sets and Calcite Cements
Five fracture sets (F1-F5) have been identified in the study area ( Figure 6) and correlated with other fractures recognized in the western part of the Sant Corneli-Bóixols anticline in previous structural studies [88,89]. We use the term "fracture set" to include both unfilled and cemented fractures (i.e., veins). This set is represented by the Bóixols thrust and minor E-W reverse faults dipping to the NNW affecting the Lluçà Formation. The Bóixols thrust is a nearly 40 km long, E-W trending, south-verging reverse fault associated with the growth of the Bóixols anticline and Coll de Nargó syncline. Although most of the Bóixols structure has been sealed by synorogenic sediments, it emerges 3 km east of the village of Bóixols and 12 km west of the Coll de Nargó ( Figure 7A). There, Jurassic dolomitic limestones from the hanging wall cut through Santonian limestones interbedded with shales of the footwall. The fault damage zone is nearly 15 m thick with a 0.5 m thick fault core, which contains dark grey-greenish gouge. The outcropping part of the thrust-related veins were sampled in order to compare them with the vein-related cements from the Coll de Nargó area.
The calcite cement Cc1 is identified in reverse faults F1 and precipitated in irregular-shaped and centimetric veins. This cement consists of: a) elongate blocky crystals with the c-axis perpendicular to the fracture walls, ranging from 0.1 to 0.15 mm wide and from 1 to 1.5 mm long ( Figure 7B), and b) anhedral to subhedral blocky calcite crystals ranging in size from 0.1 to 4 mm with dirty yellowishbrown color and thin mechanical twin planes at the microscopic scale ( Figure 7C). Most of the crystals are affected by dissolution along the edges and are crosscut by tectonic stylolites sub-parallel to the thrust trend. Elongate blocky crystals can either completely fill the fractures or partially grow from the vein edges inwards (i.e., syntaxially), where the remaining space is filled by blocky crystals. Cc1 shows a nearly black to dull-orange luminescence ( Figure 7D). This set is represented by the Bóixols thrust and minor E-W reverse faults dipping to the NNW affecting the Lluçà Formation. The Bóixols thrust is a nearly 40 km long, E-W trending, south-verging reverse fault associated with the growth of the Bóixols anticline and Coll de Nargó syncline. Although most of the Bóixols structure has been sealed by synorogenic sediments, it emerges 3 km east of the village of Bóixols and 12 km west of the Coll de Nargó (Figure 7a). There, Jurassic dolomitic limestones from the hanging wall cut through Santonian limestones interbedded with shales of the footwall. The fault damage zone is nearly 15 m thick with a 0.5 m thick fault core, which contains dark grey-greenish gouge. The outcropping part of the thrust-related veins were sampled in order to compare them with the vein-related cements from the Coll de Nargó area.
The calcite cement Cc1 is identified in reverse faults F1 and precipitated in irregular-shaped and centimetric veins. This cement consists of: (a) elongate blocky crystals with the c-axis perpendicular to the fracture walls, ranging from 0.1 to 0.15 mm wide and from 1 to 1.5 mm long (Figure 7b), and (b) anhedral to subhedral blocky calcite crystals ranging in size from 0.1 to 4 mm with dirty yellowish-brown color and thin mechanical twin planes at the microscopic scale (Figure 7c). Most of the crystals are affected by dissolution along the edges and are crosscut by tectonic stylolites sub-parallel to the thrust trend. Elongate blocky crystals can either completely fill the fractures or partially grow from the vein edges inwards (i.e., syntaxially), where the remaining space is filled by blocky crystals. Cc1 shows a nearly black to dull-orange luminescence (Figure 7d). Fracture Set 3 consists of NW-SE and NE-SW conjugate fractures, dipping between 40° and 80° and affecting only the Lluçà Formation ( Figure 8B,C). Once fracture data is rotated so that bedding is horizontal, all the data from the different outcrops fit well with a strike-slip stress field ( Figure 6). These fractures have a regular shape, straight walls and openings ranging from 1 to 15 mm. Their occurrence is concentrated in competent layers of marly limestones. Besides, some of these fractures were reactivated first as strike-slip structures and afterwards with minor dip-slip movement, as highlighted by slickenlines and overprinting striae sets on fracture planes ( Figure 8D).
F3 fractures contain two generations of calcite cements: Cc1 (described above) and Cc2, represented by a white-greyish calcite in hand sample, characterized by 0.1-0.7 mm in size elongate crystals and up to 2 mm in size blocky crystals. Evidence of dissolution along crystal edges is common. Cc2 is almost entirely non-luminescent and clearly crosscuts, and therefore postdates, Cc1 ( Figure 8E,F).  (Figure 8b,c). Once fracture data is rotated so that bedding is horizontal, all the data from the different outcrops fit well with a strike-slip stress field ( Figure 6). These fractures have a regular shape, straight walls and openings ranging from 1 to 15 mm. Their occurrence is concentrated in competent layers of marly limestones. Besides, some of these fractures were reactivated first as strike-slip structures and afterwards with minor dip-slip movement, as highlighted by slickenlines and overprinting striae sets on fracture planes (Figure 8d).
F3 fractures contain two generations of calcite cements: Cc1 (described above) and Cc2, represented by a white-greyish calcite in hand sample, characterized by 0.1-0.7 mm in size elongate crystals and up to 2 mm in size blocky crystals. Evidence of dissolution along crystal edges is common. Cc2 is almost entirely non-luminescent and clearly crosscuts, and therefore postdates, Cc1 (Figure 8e,f). Fracture Set 4 affects the entire Garumnian sequence and corresponds to NE-SW and NW-SE conjugate fractures dipping between 60° and 85°, which formed as well under a strike-slip stress field ( Figure 6). They are 20-40 cm spaced with increasing density towards fault segments of Fracture Set 5 (F5) ( Figure 9A) and have opening widths from 3 to 20 mm. Fracture edges are sharp in competent layers, such as conglomerates and lacustrine limestones, whereas in red-mudstones pressure-solution seams create fuzzy edges. Centimetric-scale vuggy porosity is associated with this type of fractures and developed by partial rock dissolution ( Figure 9C).   F4 fractures and the associated dissolution vugs are filled with two generations of calcite cements: Cc3 and Cc4. Cc3 consists of both blocky and elongate crystals growing with the c-axis perpendicular to the fracture walls and ranging from 1 to 4 mm long and 0.2 to 0.5 mm wide (Figure 10a-c) and showing an orange to bright yellow luminescence (Figure 10b). Cc4 is characterized by subhedral to anhedral blocky calcite crystals ranging in size from 0.1 to 2 mm (Figure 10c) with a dull to bright orange luminescence. Twin mechanical planes and evidence of dissolution are present in most of the borders, where spotty accumulation of iron oxides occurs. Cc4 was formed during successive deformation events as it appears extremely reworked. Some relicts of host rock fragments are also observed between Cc4 crystals (Figure 10d). F4 fractures and the associated dissolution vugs are filled with two generations of calcite cements: Cc3 and Cc4. Cc3 consists of both blocky and elongate crystals growing with the c-axis perpendicular to the fracture walls and ranging from 1 to 4 mm long and 0.2 to 0.5 mm wide ( Figure  10A-C) and showing an orange to bright yellow luminescence ( Figure 10B). Cc4 is characterized by subhedral to anhedral blocky calcite crystals ranging in size from 0.1 to 2 mm ( Figure 10C) with a dull to bright orange luminescence. Twin mechanical planes and evidence of dissolution are present in most of the borders, where spotty accumulation of iron oxides occurs. Cc4 was formed during successive deformation events as it appears extremely reworked. Some relicts of host rock fragments are also observed between Cc4 crystals ( Figure 10D).

Fracture Set 5 (F5): NW-SE and NE-SW Conjugate Strike-Slip Faults Filled with Calcite Cement Cc4
Both NW-SE right-lateral and NE-SW left-lateral conjugate strike-slip faults affect three of the studied units (Lluçà and Santa Fe Formations, and Garumnian facies), dipping from 50° to 90° both to the south and north. They intersect bedding at high angles and show moderate displacements, from centimeters up to a few meters. A right-lateral strike-slip fault (the Segre Fault) with 100 m of horizontal offset that juxtaposes the Lower Cretaceous Lluçà Formation against the Paleocene Garumnian facies was studied ( Figure 9D). In this location, deformation is mainly accommodated along a 2-3 m wide fault core. In other areas, these faults truncate and displace Mesozoic and Paleocene units to their current configuration (Figures 3 and 9E). In competent layers of the Garumnian facies, such as conglomerates and lacustrine limestones, deformation is localized on discrete planar polished slip surfaces with displacements ranging from a few centimeters to a few

Fracture Set 5 (F5): NW-SE and NE-SW Conjugate Strike-Slip Faults Filled with Calcite Cement Cc4
Both NW-SE right-lateral and NE-SW left-lateral conjugate strike-slip faults affect three of the studied units (Lluçà and Santa Fe Formations, and Garumnian facies), dipping from 50 • to 90 • both to the south and north. They intersect bedding at high angles and show moderate displacements, from centimeters up to a few meters. A right-lateral strike-slip fault (the Segre Fault) with 100 m of horizontal offset that juxtaposes the Lower Cretaceous Lluçà Formation against the Paleocene Garumnian facies was studied (Figure 9d). In this location, deformation is mainly accommodated along a 2-3 m wide fault core. In other areas, these faults truncate and displace Mesozoic and Paleocene units to their current configuration (Figures 3 and 9e). In competent layers of the Garumnian facies, such as conglomerates and lacustrine limestones, deformation is localized on discrete planar polished slip surfaces with displacements ranging from a few centimeters to a few meters. Sub-horizontal slickenlines developed on fault planes indicate a pure strike-slip motion. In incompetent red mudstone beds deformation is diffuse through centimetric shear bands with associated drag folds. In major fault segments, deformation is accommodated along a fault core up to 1 m thick, represented by a brittle incohesive fault rock formed by host rock fragments, earlier cements and fractured veins. F5 strike-slip faults are filled by calcite cement Cc4 described above.

Oxygen and Carbon Isotopes
The oxygen and carbon isotopic composition of host rocks and calcite cements are represented in Figure 11 and summarized in Table 1

Clumped Isotopes
The measured ∆ 47 values from clumped isotope thermometry of the different calcite cement generations were converted into temperatures and δ 18 O values of the precipitating fluid using the equations of [87] and [92], respectively. These values are summarized in Table 2 and Figure 12.

Strontium Isotopes
The 87 Sr/ 86 Sr ratios of host rocks and calcite cements are represented in Figure 13. Host rocks from the Lluçà Formation have a 87

Strontium Isotopes
The 87 Sr/ 86 Sr ratios of host rocks and calcite cements are represented in Figure 13

Elemental Composition
Elemental composition (major and REE + Y) was measured in the four generations of calcite cements. The main analyzed elements used for discussion are summarized in Table 3. The minimum and maximum Mn, Sr, Mg and Fe contents are also represented in Figure 14. For the complete elemental composition of the cements, see supplementary data. In the case of REE and Yttrium, four variables were used in order to better understand and constrain their variation among the different generation of calcite cements [94,95] (Figure 15): Y/Ho, Ce anomalies (Ce/Ce *), Pr anomalies (Pr/Pr *) and Log(Ce/Ce *). REE + Y have been normalized with respect to post Archean Australian Shale (PAAS) [96]. Ce and Pr anomalies were calculated following the equation of [97] and [98]. The four calcite cement generations show Y/Ho ratios ranging between 46.3 and 53.4 (Figure 15a). Cc1, Cc3 and one sample of Cc2 have positive La anomalies (field IIa), whereas Cc4 and one sample of Cc2 show negative Ce anomalies (field IIIb) (Figure 15b). and Log(Ce/Ce *) values higher than −0.10, in the case of Cc3 and some samples from Cc1 and Cc2, and lower than −0.10 in the case of Cc4 and some samples from Cc1 and Cc2 (Figure 15c).

Elemental Composition
Elemental composition (major and REE + Y) was measured in the four generations of calcite cements. The main analyzed elements used for discussion are summarized in Table 3. The minimum and maximum Mn, Sr, Mg and Fe contents are also represented in Figure 14. For the complete elemental composition of the cements, see Supplementary Data.   In the case of REE and Yttrium, four variables were used in order to better understand and constrain their variation among the different generation of calcite cements [94,95] (Figure 15): Y/Ho, Ce anomalies (Ce/Ce *), Pr anomalies (Pr/Pr *) and Log(Ce/Ce *). REE + Y have been normalized with respect to post Archean Australian Shale (PAAS) [96]. Ce and Pr anomalies were calculated following the equation of [97] and [98]. The four calcite cement generations show Y/Ho ratios ranging between 46.3 and 53.4 (Figure 15a). Cc1, Cc3 and one sample of Cc2 have positive La anomalies (field IIa), whereas Cc4 and one sample of Cc2 show negative Ce anomalies (field IIIb) (Figure 15b). and Log(Ce/Ce *) values higher than −0.10, in the case of Cc3 and some samples from Cc1 and Cc2, and lower than −0.10 in the case of Cc4 and some samples from Cc1 and Cc2 (Figure 15c).

Host Rock Diagenesis
This study analyzes fractures and calcite cements hosted in four different rock units. The Jurassic limestones, Lower Cretaceous marls and marly limestones (Lluçà Fm), and Upper Cretaceous limestones (Santa Fe Fm) belong to the hanging wall of the Bóixols thrust sheet, whereas the Paleocene continental series (Garumnian facies) belongs to the footwall. The host rocks from the hanging wall show δ 13 C values in good agreement with carbonates precipitated from Jurassic and Cretaceous seawater, but δ 18 O values are depleted with respect to marine standard values [91] (Table 1, Figure 11). This depletion in δ 18 O is interpreted to result from the increase of temperature and host rock alteration during progressive burial [103]. Finally, isotopic values from Paleocene continental facies are consistent with deposition in freshwater [104], including paleosoils and lacustrine water carbonates that contain organic matter [105].

Validity of Clumped Isotope Thermometry
The use of clumped isotopes in order to establish the temperature and δ 18 O of the fluid responsible for calcite precipitation is relatively new, and thus under validation, but it is providing promising results in the research field of diagenesis [106][107][108][109]. In this study, clumped isotopes analyses were performed in cements Cc1 to Cc4 obtaining temperatures of about 85-87 • C and δ 18 O fluid around +1‰ VSMOW in Cc1, Cc3 and Cc4, and temperatures of about 40 • C and δ 18 O fluid around −1‰ VSMOW in Cc2. Clumped isotopes can potentially be reset by recrystallization and by open-system to closed-system isotope-exchange reactions and diffusion within the mineral lattice [110,111]. In our samples, evidence of calcite recrystallization, such as grain coarsening due to grain boundary migration, is not observed. On the other hand, our vein samples have been buried at a maximum depth of 3 km during the Oligocene to the earliest Miocene, due to the deposition of syntectonic continental conglomerates [112] under a geothermal gradient about 30 • C/km, and probably lower, typical of orogenic systems. Using the parameters of [113] with the aforementioned conditions, burial temperature never exceeded 90 • C and therefore clumped isotopes could not have been modified by intra-crystal diffusion processes, as solid-state reordering requires temperatures above 100-120 • C for a time period of at least 10 million years [114]. Therefore, we can conclude that there is little chance for the clumped isotope to be reset and thus, they registered the circulation of a high-temperature fluid with a similar origin in three different stages (Cc1, Cc3 and Cc4) and a low-temperature fluid (Cc2).

Fluid-Flow Model: Fracture Development and Related Migrating Fluids
The integration of structural, petrological and geochemical analyses of calcite cements and related host rocks allows us to propose the fluid-flow model for the frontal part of the Bóixols thrust sheet from early contraction to post-folding stage.

Early Contraction Stage (T1)
During the early contraction stage (T1), the emplacement of the Bóixols thrust (F1) occurred together with the development of the two bed-perpendicular fracture sets (F2 and F3), which affect the Santa Fe and Lluçà Formations, respectively ( Figure 16). Such a synchronicity is deduced by the orientation of these fractures with respect to the shortening direction, as has also been proposed in other studies carried out in the western sector of the Bóixols anticline [89] and the presence of Cc1 in sets F1 and F3.
Calcite cement Cc1 precipitated within the Bóixols thrust (F1) and related conjugate fracture sets F3. The similarity between δ 13 C values of Cc1 (+0.4 to +1.7‰ VPDB) and host carbonates (+1.6 to +2.5‰ VPDB), indicates high fluid-rock interaction and the buffering of the carbon isotopic composition of the circulating fluid. The δ 18 O fluid calculated from clumped isotope thermometry for Cc1 (0.6 ± 0.8‰ VSMOW) falls within the range of seawater values [91] and the 87 Sr/ 86 Sr ratio (0.707468) is consistent with the values of Lower Cretaceous marine carbonates (0.707167 to 0.707492) [91,93]. In addition, the high Mg and Sr contents ( Figure 14) are consistent with the values of original seawater and the high Y/Ho ratio (>36) (Figure 15a) also supports a seawater origin but with a certain siliciclastic influence [95,115]. Moreover, the Log(Ce/Ce *) values above and below −0.10 ( Figure 15c) and the large oscillation in the Fe content may indicate fluctuation between oxic and anoxic conditions in this seawater and also supports a siliciclastic influence [116]. Additionally, the depletion in δ 18 O values of Cc1 (between −13.2 and −11.1‰VPDB) with respect to the Lower Cretaceous carbonates (−5.1 to −3‰ VPDB) ( Figure 11) is explained by the relatively high temperature calculated for this cement with clumped isotope thermometry, between 85 • C and 92 • C ( Figure 12). Therefore, although percolation of penecontemporaneous seawater through the strata at the time of vein formation cannot be discarded, the elevated temperatures obtained from geochemical analyses, together with the low abundance and small dimensions of Cc1 veins, suggest that Cc1 precipitated from connate Cretaceous seawater trapped in the host rock porosity, which was heated during progressive burial and expelled during compaction associated with the emplacement of the Bóixols thrust sheet. Such a dewatering of marine host rocks during thrusting has also been reported in other thrust faults in the Pyrenees [24]. However, one sample from the Bóixols thrust-related deformation (F1) has a considerably higher 87 Sr/ 86 Sr ratio (0.708028). This more radiogenic value could be explained by the interaction of the mineralizing fluid with Triassic sediments which constitute the detachment level of the Bóixols thrust and/or the siliciclastic fraction of marls of the Lluçà Fm and underlying formations.

Conclusions
The integration of structural, petrological and geochemical analyses performed in fracturerelated calcite cements and host rocks constrains the spatiotemporal paleo-fluid system in the frontal part of the Bóixols thrust sheet, the oldest thrust sheet of the South-Central Pyrenean Unit. Five different fracture sets were recognized affecting marine Jurassic and Cretaceous rocks from the hanging wall and continental Paleocene rocks from the footwall. These fracture sets and related calcite cements reflect the evolution of the thrust/folding system and the relationship between the tectonic evolution and fluid migration: 1) During the early contraction, two bed-perpendicular NW-SE and NE-SW fracture sets developed and were mineralized with calcite cement Cc1, which precipitated from evolved marine-derived fluids expelled from Lower Cretaceous host rocks; 2) During progressive thrusting and the main folding stage, NW-SE and NE-SW conjugate fracture sets were reopened and the Bóixols thrust was fossilized by continental Paleocene facies, allowing the infiltration of meteoric fluids that mixed with marine connate waters and precipitated calcite cement Cc2; 3) During the early post-folding stage, a NW-SE fracture set developed and calcite cement Cc3

Main Folding Stage (T2)
During the main folding stage (T2) (Figure 16), the conjugate fracture sets F3 were reactivated allowing the migration of a fluid from which calcite cement Cc2 precipitated. This is evidenced by the presence of two striae sets showing dip-slip and strike-slip kinematics (Figure 8d) and the coexistence of two calcite cements (Cc1 and Cc2) with clear cross-cutting relationships (Figure 8e,f).
The δ 18 O fluid calculated from clumped isotope thermometry for Cc2 (−1.3 ± 2.9‰ VSMOW) and its temperature (between 27 • C and 57 • C) are interpreted as the infiltration of meteoric fluids during progressive uplift, folding and fossilization of the Bóixols thrust by the continental Paleocene Garumnian facies, which mixed with previous local and connate marine waters ( Figure 16). The 87 Sr/ 86 Sr values of Cc2 (between 0.707298 and 0.707326) and δ 13 C (from +1 to +2.7‰ VPDB) are consistent with the values of Lower Cretaceous marine host carbonates. These meteoric waters lowered the Mg, Sr, Fe, and Mn contents with respect to Cc1, due to its low saturation ( Figure 14). Also, the Y/Ho, the Ce and La anomalies between IIa and IIIb fields and the log(Ce/Ce *) show a major siliciclastic signal and oxidizing-reducing transitional conditions typical of the meteoric environment.

Early Post-Folding Stage (T3)
The NE-SW and NW-SE conjugate Fracture Set F4 and its related calcite cement Cc3 developed during the latest folding stage (T3) (Figure 16). The orientation of F4 remains constant regardless of the different bedding dips, indicating that they developed once the strata had already been folded and turned vertically.
The calcite cement Cc3 shows positive δ 18 O fluid values, around +1.6 ± 0.6‰ VSMOW, which can indicate the end of the meteoric water input and again the upflowing of trapped marine waters or fluid interaction with silicate minerals. Late Cretaceous evaporated seawater is the most probable source for this cement, if we take into account that (i) the 87 Sr/ 86 Sr ratios (0.707614-0.707706) are consistent with the values of Late Cretaceous seawater [90,93], and that (ii) during the Late Cretaceous there was a transition between marine to lagoonal-lacustrine environments in the study area [77] and, in consequence, to more restricted and evaporative conditions. The depletion in δ 18 O (from −8.3 to −7.8‰ VPDB) is attributed to the relatively high temperature of the precipitating fluid (between 81 • C and 89 • C). On the other hand, the depletion in δ 13 C values (from −10.4 to −8.3‰ VPDB) with respect to typical marine carbonates [91,104], as well as the relatively low Mg and high Mn contents ( Figure 14) are explained by the combined effect of burial diagenesis and the interaction between the original seawater with Paleocene continental host rocks. Therefore, Cc3 precipitated from upflowing evolved evaporated marine fluids heated at depth and expelled from rock porosity by sediment compaction during progressive deformation and fold tightening. The Y/Ho ratio, the presence or absence of La anomaly and the Log(Ce/Ce *) values above −0.10 ( Figure 15) indicate a clear influence of siliciclastic sediments, represented by the Garumnian facies, and suboxic to anoxic paleo redox conditions [97,117].

Post-Folding Stage (T4)
During the post-folding stage (T4), two conjugate sets of NW-SE and NE-SW right-lateral and left-lateral strike-slip faults F5 developed (Figure 16), as evidenced by the constant orientation of these faults independently of the dipping of the strata. Previous studies in the Bóixols anticline also attributed these faults to the post-folding stage [89]. Strike-slip faults, observed also in other anticlines worldwide, such as in the Lurestan Province (Zagros Mountains, Iran) are developed when folds reach their maximum amplification and cannot easily accommodate any further orthogonal shortening [118]. These strike-slip faults and conjugate fractures F4 have similar orientations indicating that they formed under the same stress field ( Figure 6). As fractures F4 are crosscut by F5 or reactivated as strike-slip faults during T4, they are interpreted as a pre-slip stage of F5. This fracture pattern is similar with the "pump" model of faulting proposed by Reference [119].
The calcite cement Cc4 precipitated in F5 faults and F4 reopened fractures. This cement shows δ 18 O fluid values (+0.7 ± 0.2‰ VSMOW), δ 18 O (from −14.3 to −9.9‰ VPDB), a temperature of precipitation (around 88 • C) and 87 Sr/ 86 Sr ratios (between 0.707586 and 0.707612) similar to those obtained for Cc3, indicating that Cc4 also precipitated from heated upflowing evaporated marine fluids expelled during progressive dewatering of Cretaceous marine host carbonates during deformation in the post-folding stage ( Figure 16). The negative Ce anomaly (field IIIb) and the lowest Log(Ce/Ce *) values also supports a marine source but more oxygenated compared to earlier cements ( Figure 15) [94,95,120]. On the other hand, the Y/Ho ratios point to a variable degree of siliciclastic influence ( Figure 15). As faults F5 crosscut all the outcropping sedimentary units, from Lower Cretaceous marine to Paleocene continental host rocks, this variation may have been produced by the interaction of such fluids with marine (higher Y/Ho ratio) or continental (lower Y/Ho ratio) host rocks. Likewise, the δ 13 C (from −12.5 to −4.1‰ VPDB) values of Cc4, depleted with respect to marine carbonates [90,93], as well as the relative enrichment in Mg and Sr contents respect to Cc3, are also explained by the interaction between the upflowing evolved marine fluids with the different adjacent host rocks ( Figure 16).

Conclusions
The integration of structural, petrological and geochemical analyses performed in fracture-related calcite cements and host rocks constrains the spatiotemporal paleo-fluid system in the frontal part of the Bóixols thrust sheet, the oldest thrust sheet of the South-Central Pyrenean Unit. Five different fracture sets were recognized affecting marine Jurassic and Cretaceous rocks from the hanging wall and continental Paleocene rocks from the footwall. These fracture sets and related calcite cements reflect the evolution of the thrust/folding system and the relationship between the tectonic evolution and fluid migration: (1) During the early contraction, two bed-perpendicular NW-SE and NE-SW fracture sets developed and were mineralized with calcite cement Cc1, which precipitated from evolved marine-derived fluids expelled from Lower Cretaceous host rocks; (2) During progressive thrusting and the main folding stage, NW-SE and NE-SW conjugate fracture sets were reopened and the Bóixols thrust was fossilized by continental Paleocene facies, allowing the infiltration of meteoric fluids that mixed with marine connate waters and precipitated calcite cement Cc2; (3) During the early post-folding stage, a NW-SE fracture set developed and calcite cement Cc3 precipitated from evolved evaporated marine fluids expelled from rock porosity during progressive deformation and fold tightening; and, (4) During the post-folding stage, previous NW-SE fractures reopened, two conjugate sets of NW-SE and NE-SW strike-slip faults developed and calcite cement Cc4 precipitated again from marine evolved fluids. The comparison between the isotopic and elemental composition measured in the different fracture-related calcite cements and host rocks, reveals the progressive dewatering of Cretaceous marine host sediments during the progressive burial, deformation and fold tightening and the input of meteoric waters only during the main folding stage.