Testing the preservation potential of early diagenetic dolomites as geochemical archives

Early marine diagenetic dolomite is a rather thermodynamically‐stable carbonate phase and has potential to act as an archive of marine porewater properties. However, the variety of early to late diagenetic dolomite phases that can coexist within a single sample can result in extensive complexity. Here, the archive potential of early marine dolomites exposed to extreme post‐depositional processes is tested using various types of analyses, including: petrography, fluid inclusion data, stable δ13C and δ18O isotopes, 87Sr/86Sr ratios, and U‐Pb age dating of various dolomite phases. In this example, a Triassic carbonate platform was dissected and overprinted (diagenetic temperatures of 50 to 430°C) in a strike‐slip zone in Southern Spain. Eight episodes of dolomitization, a dolostone cataclasite and late stage meteoric/vadose cementation were recognized. The following processes were found to be diagenetically relevant: (i) protolith deposition and fabric‐preservation, and marine dolomitization of precursor aragonite and calcite during the Middle–Late Triassic; (ii) intermediate burial and formation of zebra saddle dolomite and precipitation of various dolomite cements in a Proto‐Atlantic opening stress regime (T ca 250°C) during the Early–Middle Jurassic; (iii) dolomite cement precipitation during early Alpine tectonism, rapid burial to ca 15 km, and high‐grade anchizone overprint during Alpine tectonic evolution in the Early Eocene to Early Miocene; (iv) brecciation of dolostones to cataclasite during the onset of the Carboneras Fault Zone activity during the Middle Miocene; and (v) late‐stage regression and subsequent meteoric overprint. Data shown here document that, under favourable conditions, early diagenetic marine dolomites and their archive data may resist petrographic and geochemical resetting over time intervals of 108 or more years. Evidence for this preservation includes preserved Late Triassic seawater δ13CDIC values and primary fluid inclusion data. Data also indicate that oversimplified statements based on bulk data from other petrographically‐complex dolomite archives must be considered with caution.


INTRODUCTION
Stoichiometric dolomites (dolostones) are the most thermodynamically-stable carbonate phase, surpassing even low-Mg calcites, which have been referred to as the most reliable archives of their depositional environment (Nordeng & Sibley, 1994). This also holds true with respect to earliest diagenetic marine dolomites. Previous studies have exploited these carbonates as archives of their palaeo-porewater properties (Geske et al., 2012), often with reference to replacement dolostones or primary dolomite precipitates in modern sabkha and peritidal-supratidal settings (Land, 1980;Patterson & Kinsman, 1982) or in various ancient carbonate deposits (Mazumdar & Strauss, 2006;Johnston et al., 2010;Kasemann et al., 2014). These authors argued that fabric-preserving, early diagenetic marine dolomites may largely escape subsequent alteration. Considering the mineral chemistry and temperature-dependent fractionation patterns between fluids and minerals, early diagenetic dolomites offer insight into time-resolved patterns in marine dynamics (Halverson et al., 2005;Geske et al., 2012). Other research questions included the redox state of past oceans (Kamber & Webb, 2001;Hood et al., 2018), element fluxes between oceans and continents (Tipper et al., 2006), and the geochemical cycles of C, Ca and Mg through time (Arvidson et al., 2006;Farka s et al., 2007;Arvidson et al., 2011).
The aims of this paper are three-fold; (i) to identify the complex diagenetic pathways of allochthonous Triassic dolostones in the Neogene Carboneras Fault Zone in Southern Spain by mineralogy, petrology and (isotope) geochemistry; (ii) to place these findings into context of hydrothermal fluid properties reconstructed from fluid inclusion data and geochemical fingerprints; and (iii) to test the resistance of a suite of early diagenetic dolostones to mechanical and hydrothermal alteration culminating in cataclastic fault gouge formation. This study provides a basis for the reconstruction of postsedimentary dolostone diagenesis that may be linked to discrete tectonic events, palaeoenvironmental changes, subsurface fluid chemistry and circulation patterns. These data also suggest limitations concerning the value of bulk dolostone chemostratigraphy in interpreting diagenetic histories.
Formation of the Carboneras Nijar Basin (Figs 1 and 2A) was initiated by regional subsidence and subsequent marine ingression. The lower Serravallian/early Tortonian marls of the Saltador Formation represent the oldest marine sediments (Van de Poel, 1992;Fortuin & Krijgsman, 2003). The Saltador is overlain unconformably to the north by the fossiliferous calcarenite and marls of the Azagador Formation (V€ olk & Rondeel, 1964;Fortuin & Krijgsman, 2003), which rest directly on basement rocks (Figs 2A, 3 and 4). The Azagador is in turn overlain unconformably by three Messinian (Neogene) formations: The Abad (marls), the Yesares (Messinian Salinity Crisis gypsum) and the Feos (marls; Fortuin & Krijgsman, 2003). These Messinian deposits are overlain by Pliocene (Neogene) horizons, which are inclined towards the south of the Sierra Cabrera. The younger units are progressively less inclined, indicating progressive uplift (Mart ın et al., 2003). Following the predominance of marine conditions during the Serravallian to Pliocene, the Pleistocene glaciation intervals resulted in drying of the basin and  Braga, 1987). Note that the Alpujarride carbonates occur exclusively as faulted blocks in the study area. Therefore, no assignment to specific formations was carried out, and all stratigraphical ranges in the figures are given accordingly for the Anisian-Rhaetian (Middle/Late Triassic) time interval. Refer to Mart ın & Braga (1987) for details and depositional settings of the individual facies types 1 to 9. LST, limestone.
subaerial exhumation of the fault zone. For additional information about the regional geology see Rutter et al. (2012). A summary of the Neogene to Recent settings in the study area is shown in Fig. 3.

Sampling strategy
The Alpujarride dolostones were chosen for this study due to their abundance and diverse range of hydrothermal and tectonic overprint in a well-exposed 'natural laboratory', in addition to their well-established regional and tectonic framework (Rutter et al., 2012, 2014, andreferences therein). Fieldwork was mainly performed in the northern strand of the Carboneras Fault Zone (Fig. 4)   The polarized images were taken on a Leica DM4500P microscope (Leica Microsystems GmbH, Wetzlar, Germany). Cathodoluminescence analyses were performed using a hot cathode (HC1-LM) facility developed at the Ruhr-University Bochum (Neuser et al., 1996) equipped with a DC73 video camera system (Olympus, Hamburg, Germany). The thin sections were sputter coated with a 15 nm thick gold layer to avoid charging. The electron beam had an acceleration voltage of the 14 kV, current density from 5 to 10 lA mm À2 , and beam current between 0Á1 mA and 0Á2 mA. Microthermometric analyses were performed using a Linkam THMS600 stage (Linkam Scientific, Tadworth, UK) at the University of T€ ubingen, Germany. Fluid petrographic analyses were performed on doubly-polished thick sections (100 lm). For each fluid inclusion, the assemblage petrography was carefully documented by optical microscopy (Goldstein & Reynolds, 1994). These fluid inclusion assemblages (FIA) were paragenetically classified as primary (p), pseudo-secondary (ps), secondary (s), isolated inclusions (iso) or clusters of inclusions (refer to Walter et al., 2015, for details). For each analysis, a thrice-repeated heating and freezing cycle was conducted to receive the final dissolution temperature of ice (T m,ice ) and hydrohalite (T m,hh ), and the uncorrected homogenization temperature (T h ). For interpretation only, fluid inclusions were used with a minor variability below 0Á1°C for T m,ice and T m,hh , and below 1°C . The map was generated using the data of Rutter et al., 2012, and complemented (in the northern part) with data generated in this study. It shows the Neogene-Quarternary deposits of the Carboneras Nijar Basin and underlying basement rocks being faulted by the Carboneras Fault Zone. The study area is outlined in blue, and the location of the schematic section in Fig. 3 is indicated as A to A'. Formation names are based on those used by Van de Poel (1992), Fortuin & Krijgsman (2003) and Rutter et al. (2012), with some modification.  The salinity and molar Ca/(Na + Ca) ratios in the ternary NaCl-CaCl 2 -H 2 O system were quantified by using the MS EXCEL©-based calculation sheet of Steele-MacInnis et al. (2011). The volume fraction of each fluid inclusion was estimated based on published volume proportion tables (e.g. Shepherd et al., 1985). The data is presented in volume fraction notation, which illustrates the phase assemblage at room temperature [Lx, numerical subscription refers to the volume percentage of aqueous liquid, vapor (Vx) and solid (Sx)] (Shepherd et al., 1985;Bakker & Diamond, 2006). Pressure correction of the T h values was quantified using the HOKIEFLINC_-NACL_H2O program of Steele-MacInnis et al.

Uranium-lead dating of dolomite phases
Four dolomite and two calcite phases were analyzed using the U-Pb small scale isochron dating (UPbSSI) method. The data were acquired from polished sections by laser ablationsector fieldinductively coupled plasma-mass spectrometry (LA-SF-ICPMS) at the Goethe University Frankfurt. A Thermo Scientific Element 2 sector field ICP-MS (Thermo Fisher Scientific, Waltham, MA, USA) was coupled to a RESOLution S-155 (Resonetics, Nashua, NH, USA) 193 nm ArF Excimer laser (CompexPro 102; Coherent, Palo Alto, CA, USA) equipped with a two-volume ablation cell (Laurin Technic, Canberra, Australia). Samples were ablated in a helium atmosphere (0Á3 l min À1 ) and mixed in the ablation funnel with 0Á9 l min À1 of argon and 0Á06 l min À1 of nitrogen. Static ablation used a spot size of 213 lm and a fluence of ca 1 J cm À2 at 15 Hz. Each analysis consisted of 20 sec of background acquisition, followed by 20 sec of sample ablation and 25 sec of washout. Prior to analysis, each spot was pre-ablated for 5 sec to remove any surface contamination. Soda-lime glass SRM-NIST614 was analyzed together with two carbonate reference materials to bracket sample analysis. Analysis spots were set carefully on carbonate domains appearing homogeneous in transmitted light microcopy and on cathodoluminescence images, to avoid or minimize the ablation of contaminants such as clay minerals of detrital origin.
Raw data were corrected offline using a macrobased in-house MS EXCEL© spreadsheet program (Gerdes & Zeh, 2009). Data were plotted in the Tera-Wasserburg diagram and ages calculated as lower intercepts using ISOPLOT 3.71 (Ludwig, 2007). Points of each sample data set derived from a small area (<1 cm 2 ) and defining linear arrays in the 207 Pb/ 206 Pb versus 238 U/ 206 Pb space (Tera-Wasserburg diagram, Tera & Wasserburg, 1972) were interpreted to be a mixture of initial common Pb and Pb that formed due to in situ decay of U since mineralization. The age of formation is defined by the lower intercept with the Concordia. See Appendix A1 for a detailed description of the UPbSSI method.

Geochemical analyses
A total of 111 subsamples was extracted using a hand held drill. Those with a mixture of calcite and dolomite were treated with 0Á27 m Di-Na-EDTA to dissolve the calcitic phase (Geske et al., 2015b).
The carbon and oxygen isotope values (d 13 C and d 18 O) were measured by reacting 0Á1 AE 0Á01 mg of sample powder with phosphoric acid at a temperature of 70°C. Sample aliquots were measured on a Thermo Finnigan MAT 253 mass spectrometer equipped with a Gasbench II (Thermo Fisher Scientific) and a GC PAL auto sampler (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 0Á08& and 0Á06&, respectively for d 13 C, and 0Á13& and 0Á10&, respectively for d 18 O. All values are reported in & Vienna-Pee Dee Formation belemnite (V-PDB) and errors as 1r SD.
Samples were analyzed for 87 Sr/ 86 Sr ratios by dissolving a powdered sample containing ca 400 ng of Sr (<1 to 10 mg of powder) with 1 ml of 6 M HCl for about 24 h at room temperature. Samples were then dried on a hot plate at 125°C and re-dissolved in 0Á4 ml of 3 M HNO 3 . The Sr fraction was separated from the matrix using PFA columns filled with TRISKEM ion exchange resin. The samples were flushed with 1 ml of 0Á05 M HNO 3 and conditioned with 3 ml of 3 M HNO 3 to remove the matrix. The Sr fraction was then recovered with 2 ml of deionized H 2 O. The samples were evaporated on a hot plate at 90°C and then re-dissolved in 1 ml of H 2 O 2 -HNO 3 (1 : 1) to remove organic remains. After evaporating the samples again on a hot plate at 60°C and re-dissolving in 0Á4 ml of 6 N HCl, they were dried at 90°C. Finally, samples were re-dissolved in 1 ll of ionization-enhancing solution (after Birck, 1986) and loaded on Re single filaments. The loading, column and reagent blanks were <5 pg, <1 ng, and <0Á01 ppb, respectively. The 87 Sr/ 86 Sr ratios were analyzed on a Thermal Ionisation Mass Spectrometer (TIMS) TI-Box (formerly MAT 262; company -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 0Á710241 AE 0Á000002 2r SE and 0Á000032 2r SD (n = 431) for NIST NBS 987, and a 87 Sr/ 86 Sr ratio of 0Á709159 AE 0Á000002 2r SE and 0Á000031 2r SD (n = 377) for USGS EN-1. Errors of 87 Sr/ 86 Sr ratios are reported in 2r SE. A similar analytical technique is described by Faure & Powell (1972).

RESULTS
Based on field observations, three types of Alpujarride dolostones, from least to most altered, were recognized: (i) Intact, finely (millimetresized) laminated, early diagenetic dolomitized grainstones with preserved lamination, ripple foresets and fining-upward cycles (Fig. 5B). Dolo-grainstones are also characterized by blocky zebra saddle dolomite (Fig. 5C), which formed bedding-parallel or along cleavage planes, and are typically accompanied by stylolites. Zebra dolomites were recently interpreted to form during burial by dissolution-void creation within sedimentary laminations, resulting in an 'ABBA' sequence (Arecrystallized host rock; Bwhite saddle dolomite; Vandeginste et al., 2005). Zebra dolomite is common in the Alpujarride dolostones, and known from other localities in the Betic Cordilleras (e.g. Sanz de Galdeano & L opez Garrido, 2014). For more details on zebra dolomite fabrics see Vandeginste et al. (2005), Morrow (2014) or , and references therein. (ii) Partially to fully brecciated dolomitized grainstones with variable degrees of clast fitting (Fig. 5D). Carbonate cements locally form the host carbonate in which clasts are embedded. Bedding features may be preserved within individual clasts but are commonly broken up, and the space between clasts is occluded by a matrix of dolomitic cataclasite. (iii) Dolomitic cataclasite (i.e. fault gouge), with millimetre to centimetre-sized dolostone clasts (Fig. 5E). All sedimentary features are overprinted and disintegrated. Evidence of hydrothermal alteration of the cataclasite facies includes its brown-yellow colouration. In the north-western study area, dolostones are present as strongly-altered ironmanganese ores, previously mined over the last century.

Petrography and paragenetic succession
A paragenetic succession of the Alpujarride dolostones has been established from field observations (Fig. 5), thin section data and cathodoluminescence analysis (Figs 6 and 7; Table 1). Only the most significant features are reported here, but additional material can be found in the Appendix A1. Dolomite terminology is after Sibley & Gregg (1987). Eight dolomite cements are present: Dol 1, recrystallized early diagenetic, fine to coarse crystalline planar-s dolomite (Figs 6A to D and 7); Dol 2, saddle dolomite cements characterized by curved crystal surfaces and sweeping extinction forming coarse crystalline nonplanar dolomite (Figs 6 and 7A to D); Dol 3, volumetrically less significant fine to medium crystalline nonplanar dolomite ( Fig. 6E and F); Dol 4, fine crystalline nonplanar dolomite (Figs 6A, 6B, 6E to H, 7A and 7B); Dol 5, fine to coarse crystalline nonplanar to planar-s dolomite (Figs 6 and 7); Dol 6, fine crystalline nonplanar dolomite (Appendix A1; Fig. A1E and F); Dol 7, fine to coarse crystalline nonplanar to planar-s dolomite (Figs 6E, 6F and 7); and Dol 8, medium to coarse crystalline nonplanar dolomite ( Fig. 6E to H). The paragenetically-youngest dolomitic material is present in the form of a cataclasite (C 1) ( Fig. 7G and H), which does not qualify as a diagenetic phase, but as a tectonic product. This cataclasite consists of crushed (micronsized) crystals of the eight dolomite phases. Latest stage calcitic cements (LMC 1 to LMC 4, Figs 6E to H, 7C and 7D; and HMC 1 to HMC 3, Fig. 6E to H) and authigenic quartz (characterized by intrinsic dark blue luminescence) do not occur in a defined order of precipitation, but were placed into context based on appearance. In contrast, Malaguide dolostones generally consist of fine to medium crystalline planar-e to planar-s dolomite containing wellpreserved and recrystallized ooids.
Two sets of stylolites occur in the Alpujarride dolostones: Type 1 stylolites (S1) formed between layers of different grain size in the Dol 1 early diagenetic dolostone, and are often cut by paragenetically younger Dol 5 or Dol 7 cements ( Fig. 7A and B); Type 2 stylolites (S2) formed after the Dol 1 early diagenetic dolostone and the Dol 2 zebra saddle dolomite, and are cut by veins infilled by Dol 5 or Dol 7 cements ( Fig. 6C and D). The dolostones have a low porosity, as most pore space has been occluded by dolomite and calcite cements.

Microthermometry
A classification was made to place the fluid inclusions into temporal context ( Fig. 8A to C). Primary inclusions with irregular and angular shapes with dimensions of <10 lm occur along crystal growth zones. Secondary fluid inclusions post-date the whole paragenetic assemblage, whereas pseudo-secondary inclusions are related to healed cracks and often pre-date primary fluid inclusions of the next dolomite generation.
Two types of fluids were recognized (Figs 8 and 9): The first fluid type occurs in primary inclusions and fluid inclusion clusters, with both inclusion types common in Dol 1, 2, 4, 5, 7 and 8. These record the first melting point in the binary NaCl-H 2 O system, with a eutectic temperature at À21Á2°C (Fig. 9, eutectic 2). Ice occurs as the last dissolving phase whereas hydrohalite dissolves at the binary eutectic. The calculated final melting temperature of ice is between À2Á8°C and À19Á3°C, resulting in calculated salinities from 4Á7 to 21Á9 wt.% NaCl eq . Data are scattered over a wide range, but if looking at a specific position in the paragenetic sequence, defined clusters are observed. With progressing temporal evolution, the salinity of diagenetic fluids increases. Uncorrected homogenization temperatures vary between 45°C and 230°C, and also increase within the paragenetic sequence. Within a defined fluid inclusion assemblage, salinity and T h are nearly constant, but typically vary between assemblages. The inclusions contain volume fractions of ca L 90 V 10 to L 95 V 5 .
The second fluid type (only in sdol 8) is classified as H 2 O-NaCl-CaCl 2 fluid with a characteristic first melting point around the ternary eutectic at À52Á0°C (Fig. 9, eutectic 1). These inclusions are present in secondary cavities of healed fractures in dolomite phase Dol 8. The last melting temperature of ice is between À14Á1°C and À20Á3°C, whereas hydrohalite dissolves between À24Á0°C and À26Á0°C, resulting in a salinity of 21Á1 to 22Á0 wt.% NaCl + CaCl 2 . The inclusions contain volume fractions of about L 90 V 10 . Micro Raman Spectroscopy analyses of this type failed due to the subcriticallysmall size of the inclusions.

Uranium-lead age data of dolomite phases
Dolomite phases range in 207 Pb/ 206 Pb values from 0Á541 to 0Á835, and in 238 U/ 206 Pb values from 0Á39 to 13Á57. In the Tera-Wasserburg diagram, this results in lower intercept ages of 175Á0 AE 11Á0 Ma, 186Á1 AE 8Á0 Ma and 186Á4 AE 5Á6 Ma using initial 207 Pb/ 206 Pb values of 0Á841 to 0Á850 (Fig. 10). Based on scanning electron, cathodoluminescence and transmitted light microscopy, no evidence for uranium minerals was recognized in any of the samples. The U-Pb dolomite ages record recrystallization and mineralization between the Early and Middle Jurassic (164Á0 to 194Á1 Ma). The ages of both early diagenetic Dol 1 and zebra saddle Dol 2 were indistinguishable. Cement phases Dol 5 and Dol 7, and LMC 2/LMC 3 were not datable due to low U concentration and a Pb concentration that is too high, respectively.  and À0Á4& (Neogene bivalve, n = 2).

INTERPRETATION AND DISCUSSION
The paragenetic succession of the dolostones is presented and discussed below within genetic and temporal context. The interpretation and discussion of these rocks commences with the deposition of Anisian-Rhaetian (247 to 201 Ma) precursor carbonates in a platform environment, and culminates in Holocene meteoric diagenesis of the exhumed Carboneras Fault Zone and the related lithologies.

Evolution of a complex dolostone archive
Protolith deposition and early marine diagenetic dolomitization The precursor carbonates of the dolostones were deposited on a large (ca 400 km in an east-west direction) Anisian-Rhaetian carbonate platform, Fig. 11. Complete paragenetic sequence of diagenetic phases with typical cathodoluminescence colours related to their diagenetic and tectonic environment of precipitation or formation. Remaining pore space is marked in black.
likely belonging to the opening Proto-Atlantic region (Delgado et al., 1981;Mart ın & Braga, 1987). Well-preserved sedimentary features (Fig. 12) of the precursor laminar tidal flat facies (Mart ın & Braga, 1987) include flaser bedding and ripple cross-lamination ( Fig. 12A to C). This preservation suggests an early diagenetic, fabricpreserving replacement dolomitization. Low reconstructed fluid temperatures (median of 50°C) in preserved dolomite rhombs from the later recrystallized Dol 1 further imply shallow burial dolomitization. Fluid inclusion salinity of 5 wt.% NaCl, in the absence of CaCl 2 , indicates a modified seawater parent fluid (i.e. moderately saline brines) (Fig. 8C, D and F). Dolomitized Middle Triassic barrier-island facies to the west of the study area were similarly interpreted as the product of early reflux dolomitization (Mart ın, 1978). This interpretation is strengthened by the absence of calcite cement formation, possibly due to previous detachment of Ca-ions during ordering of the Dol 1 crystal structure driven by tidal pumping or convection cells in the platform. Carbon isotope (d 13 C) values of fine crystalline dolomicrite Dol 1 (0Á7 to 1Á8&) remain within the range of their Middle/Late Triassic marine formation fluids (ca 0 to 4&) (Fig. 13). In contrast, oxygen isotope (d 18 O) data of Dol 1 were shifted towards more negative values (À6Á5 to À8Á0&), and therefore do not preserve the early diagenetic dolomitization signature of Middle/ Late Triassic marine fluids (ca À6 to 0&). Fluid inclusion data suggest that strongly evaporative marine porewaters, typically resulting in a positive shift in d 18 O values, were not present ( Fig. 8A to E). Fractionation during LST to dolostone neomorphosis (at Earth surface temperatures) is also expected to cause a positive shift in d 18 O values (Swart, 2015). Cathodoluminescence data help to explain the unexpectedly low d 18 O values observed. The patchy luminescence of Dol 1 indicates partial recrystallization of the early diagenetic dolomite phase (Fig. 11). Numerous bright luminescent fissures cross-cut the Dol 1 matrix and suggest influence of a significantly deeper burial fluid (Fig. 6B). This fluid-rock interaction may have resulted in the resetting of the fluid-buffered oxygen isotope, but not the rock-buffered carbon isotopes (Fig. 13). Furthermore, a different mechanism to explain the lowering of d 18 O values could be early meteoric overprint of the dolostones (Lohmann, 1987); however, this seems unlikely due to the absence of meteoric vadose or phreatic textures. homogenization temperatures and salinities between Dol 1 and Dol 2 (45 to 52°C versus 100 to 150°C, and 5 versus 11 wt.% NaCl, respectively; Figs 8 and 9) support a different scenario. These data suggest the precipitation of Dol 1 in early diagenetic shallow marine burial conditions, and a significantly later (tens of Myr) hydrothermal precipitation of Dol 2. Along with the structural evidence and higher salinity, fluctuating fluid inclusion temperatures are characteristic for hydrothermal dolomite cements including zebra dolomite (Wilson et al., 1990;Morrow, 2014), further supporting the hypothesis of different formation fluids of Dol 1 and Dol 2. Early diagenetic Dol 1 recrystallized, and therefore bears similar d 13 C and d 18 O values as Dol 2. However, some of the primary fluid inclusions from Dol 1 remained intact in non-recrystallized dolomite rhombs, preserving the initial signature of Dol 1 fluids. Precipitation of the zebra saddle dolomite (Dol 2) may have occurred after the recrystallization of Dol 1, as S1 stylolites postdate the recrystallization, and are arguably coeval with the opening of voids in which the saddle dolomite (Dol 2) nucleated ( Fig. 7A and B).

Intermediate burial and hot hydrothermal overprint
Uranium-lead (U-Pb) age data (164Á0 to 194Á1 Ma including errors) for Dol 1 and Dol 2 coincide with Early to Middle Jurassic tectonism that induced far-field hydrothermal pulses recorded across Europe (Walter et al., 2018, and references therein). During several tectonic events, large portions of the upper crust were saturated with hot fluids through extensional basinrelated fault systems, causing alteration of the basement and mineralization (Tritlla & Sole, 1999;Meyer et al., 2000;Cathelineau et al., 2012;Subias et al., 2015;Munoz et al., 2016). The dominant tectonic event was the rotation of Iberia, caused by the opening of the Proto-Atlantic (e.g. Wetzel et al., 2003;Clauer et al., 2008;Pfaff et al., 2009;Brockamp et al., 2011;Walter et al., 2017). Most studies have primarily focused on the mineralization of clastic and magmatic Palaeozoic to Triassic rocks. Possibly for the first time, this large-scale hydrothermal event is now documented in carbonates. It seems likely that similar features are recorded across the Iberian region, but most studies lack constraining age data.
Despite many hydrothermal dolomite deposits (e.g. Davies & Smith, 2006;Jacquemyn et al., 2014;Morrow, 2014) developing as replacement minerals of 'pure' LST precursors, the example documented here arguably evolved from a dolomite precursor, because only dolomite rhombs were preserved after the recrystallization of Dol 1.
Although the iron-rich Dol 3 is less common than Dol 1 and Dol 2, its partial corrosion ( Fig. 6E and F) indicates a change in fluid chemistry and, likely, a precipitation hiatus. The subsequent Dol 4 and Dol 5 phases are characterized by fluid inclusion homogenization temperatures up to 250°C (medians of 215°C and 235°C, respectively; Fig. 8B, D and E) suggesting burial depths in excess of some thousands of metres (Davies & Smith, 2006;Hu et al., 2018). Elevated fluid inclusion salinities (17 to 20 wt.% NaCl typical for basinal or continental basement brines; Frape et al., 1984) distinguished Dol 4 and Dol 5 from parageneticallyolder zebra saddle dolomite formed at shallower burial depths (Fig. 8F). A structural (morphological) roughness analysis (Renard et al., 2004) of five bedding-parallel stylolites of the studied dolostone was calculated from the Fourier Power Spectra P(k) to estimate: (i) the compressive stress direction (Koehn et al., 2012); (ii) the depth of formation (Bertotti et al., 2017); and (iii) the amount of compaction (Ebner et al., 2009b). Stylolite groups S1 and S2 (Figs 6C, 6D, 7A and 7B) formed under the same compactiondriven stress regime, at a calculated formation depth of 400 to 800 m and pressure of 10 to 18 MPa. These calculations do not imply high fluid temperatures related to deep burial. The S2 stylolites post-date Dol 5 ( Fig. 6C and D) and, therefore, pre-date Alpine orogenesis, suggesting that bedding was still oriented horizontally when cement phases Dol 4 and Dol 5 formed.
Following Ingebritsen & Sanford (1998), riftrelated geothermal convection is a possible mechanism to bring hot, deep-seated fluids into shallower burial environments. This is a likely explanation here, due to the short-lived pulses of hydrothermal fluids along extensional faults during the opening of the Proto-Atlantic in the Early Jurassic to Early Cretaceous. Braithwaite et al. (2004) discuss the common occurrence of deep seated thermobaric (Davies & Smith, 2006) fluid injection into the compactional regime. A significant implication of these data is that the dolostones remained within the same compaction-driven stress regime for nearly half of their diagenetic history (ca 100 Myr).
Hydrothermal and multistage tectonic overprint during the Alpine Orogeny Following the precipitation of Dol 5, the contribution of hydrothermal fluids continuously decreased along the later diagenetic phases, becoming least significant in dolomite cement phases 7 and 8 (215 to 250°C in Dol 5 versus 150 to 165°C in Dol 8; Fig. 8B, D and E). A high amount of fluid inclusions and the scattered occurrence of Dol 6 indicate significantly reduced precipitation rates (possibly a hiatus) with high fluid temperatures. Dol 7 and Dol 8 are characterized by similar fluid inclusion salinities relative to those of Dol 4 and Dol 5 (17 to 22 wt.% versus 17Á5 to 20Á5 wt.% NaCl, respectively; Fig. 8C and F). However, fluctuating fluid inclusion densities (1Á02 to 1Á10 g cm À3 ; Fig. 8F) and lower homogenization temperatures (median of 155 to 165°C; Fig. 8B, D and E) help to differentiate Dol 7 and Dol 8 from Dol 4 and Dol 5 (Fig. 8B, D and E). The change in Dol 8 inclusions towards CaCl 2 -fluid chemistries (Fig. 9) evidences dissolution of evaporites and hence a different fluid source. Moreover, Dol 7 heralds the onset of a different tectonic setting than the previous compactional regime. Thin section evidence indicates that Dol 7 and Dol 8 phases occlude pore space in tectonic dolostone breccia (Fig. 7C to F), which is attributed to early Alpine thrusting (Lonergan & Platt, 1995;Platt et al., 2005). Dolomite cements 7 and 8 yield comparable d 18 O values relative to the paragenetically-older phases, suggesting fluid-controlled homogenization (Fig. 13). Petrographic and geochemical data indicate that the homogenization event may be related to late-stage dolomite cementation. In contrast, Dol 7 and 8 are moderately 13 Cdepleted (d 13 C min. = À1Á7&, Table 1) relative to older precipitates. The patterns observed here are in good agreement with data from other structurally-controlled hydrothermal dolomites worldwide (Davies & Smith, 2006).
Arguments supporting Alpine orogenesis as the driver of large-scale fluid movement and formation of dolomite cements 7 and 8 include: (i) relative age constraints on the complex vein dolomite phases from cathodoluminescence analyses (Fig. 7); and (ii) the relation of these pore-filling fabrics in tectonic breccias of Alpine age that predate the Carboneras fault activity (as supported by previous regional studies). More specifically, contractional tectonism with Ar-Ar ages suggests Alpine Orogeny between the Eocene and Early Miocene (Monie et al., 1991;Platt et al., 2005). An initial phase of this major tectonic phase could likely be a trigger for the expulsion of hot, deep-seated saline continental basement brines (Fig. 9) that induced the precipitation of dolomite cements 7 and 8. Similar patterns have been reported in several earlier studies worldwide (i.e. Davies & Smith, 2006;L opez-Horgue et al., 2010;Hendry et al., 2015;Mart ın-Mart ın et al., 2015;Dong et al., 2017). A second, large-scale tectonic event is arguably recorded here, the modification of the European Cenozoic Rift System (Maillard & Mauffret, 1999;Walter et al., 2018, and references therein) prior to, and during, the Alpine Orogeny. This stage may also have reactivated Jurassic to Cenozoic extensional faults, resulting in a proto-Carboneras fault system. This progression of tectonic events was followed by accelerated burial in the Oligocene that caused high temperature and high pressure metamorphosis. During the Early Miocene, rapid exhumation of large areas of the Betic Cordilleras occurred (Vissers, 2012). Substantial evidence exists for greenschist facies metamorphism in bulk Alpujarride strata from analyses in metapelites, phyllites and calcareous/dolomitic 'marbles' (see e.g. Bakker et al., 1989;Azañ on & Crespo-Blanc, 2000;Vissers, 2012; and many others). Remarkably, late-stage, lowgrade metamorphism with temperatures between 300°C and 430°C, did not transform the Carboneras dolostones into marble, as evidenced by the (albeit patchy) preservation of dolomite cements and their primary fluid inclusions (Figs 6 to 9).
Discrepancy between studies regarding burial depth of the dolostones analyzed here merits attention. According to Bakker et al. (1989) and Vissers (2012), the Alpujarride strata in the eastern Betics were buried to depths of 25 km (P: 700 MPa; T: 430°C) during the Eocene to Oligocene. These authors propose a reduced geothermal gradient of 17 to 20°C rather than 30°C per kilometre. Assuming that these assumptions are true, two possible interpretations exist: (i) Alpujarride dolostones were indeed buried to depths of 25 km and temperatures in excess of 400°C, but preserved their structural integrity, petrographic characteristics and even primary fluid inclusions; or (ii) the Alpujarride dolostones studied here experienced P/T pathways and burial histories that differed from other Alpujarride rock units. At present, neither of these interpretations are conclusive. Based on the presented evidence, it seems likely that deep burial over time spans of at least 20 to 25 Myr (Eocene to Oligocene) should have resulted in the neomorphosis of minerals typical of these conditions, such as periclase, brucite, phyllosilicates or serpentine (Ferry et al., 2002;Bucher & Grapes, 2011), which are not found in the described units.
Moreover, even in an end-member type, rockbuffered system, carbonates are expected to recrystallize, albeit very slowly. This would occur in a diffusion-controlled system, and in the presence of thin fluid films at crystal boundaries or along reaction fronts (Pingitore, 1982;Watson & M€ uller, 2009;Jonas et al., 2017;Pederson et al., 2019). These slow, fabricpreserving micro-scale dissolution-recrystallization processes are probably under-appreciated and under-explored from a carbonate archive perspective. Clearly, Alpujarride dolostones experienced physical and chemical compaction during burial, and Dol 2 to Dol 5 cements occluded the remaining pore space, resulting in fabric that was generally impermeable for fluids. However, even at great depths, spatiallylocalized (overpressured) pathways for aqueous fluids exist along fractures and fault systems, often resulting in complex diagenetic systems. Considering the data discussed here, it seems likely that these dolostones experienced peak burial depths of ca 15 km and temperatures of up to 300°C. Deeply-buried, low porosity dolostonesnot yet transformed to marbleshave been reported from other low-grade metamorphic zones in the Alpine orogen. For example, low porosity, Upper Triassic to Lower Jurassic dolostones in the Alpi Apuane (north-west Italy), preserved their textures under temperatures up to 300°C, and only then did recrystallization commence (Cortecci et al., 1999;Molli et al., 2018, and references therein).
In summary, the dolostones comprising the uppermost units of the Alpujarride preserved most of their early diagenetic textures even under low-grade metamorphic conditions or, alternatively, underwent fabric-preserving micro-scale recrystallization. However, the observation of primary fluid-inclusions in these rocks does not support the latter scenario. Timing of the homogenization of dolomite d 18 O signatures of all cement phases pre-dating (and potentially including) Dol 7 and Dol 8 remains speculative. Considering the previously mentioned rock-buffered and fluid-lean system, extended durations (Early Eocene to Early Miocene) and temperatures ca 300°C would be required for the slow diffusion processes to occur across thin fluid films at crystal boundaries to facilitate d 18 O homogenization. These processes are poorly understood from a mechanistic and kinetic perspective, and further wellcalibrated experimental work is needed.

Tectonic and renewed hydrothermal overprint in the Carboneras Fault Zone
Evidence for the overprint of fault zone rocks comes (among other aspects) from the relationship between allochthonous dolostones and autochthonous transgressive deposits of the Carboneras Nijar Basin. The Azagador Formation with marine invertebrate remains (Fig. 3) evidences the flooding of the fault zone during the Late Tortonian. Although no direct evidence exists for the timing of subsequent events that led to varying degrees of cataclasation in the fault zone (Fig. 14), the position of dolostone blocks in the Fault Zone and the described paragenetic sequence suggest that this process occurred during the main phase of tectonic activity (Middle to Late Serravallian).
Based on dolomite geochemistry, an increasing degree of brecciation and cataclasation (Figs 11 and 13) coincides with a D 18 O of 4& between early marine dolomicrite Dol 1 and cataclasite C 1 (Fig. 13). Importantly, not only the cataclastic matrix yields 18 O-enriched values compared to the 'precursor' phases, but a resetting of the Dol 1 clasts within this breccia is also observed. This indicates an exchange between fault zone fluids and rock fragments, potentially due to a newly increased permeability. The fluid-dominated fault system and the large surface area to volume ratio of clasts facilitated the resetting of bulk d 18 O signatures. The shift to higher d 18 O valuesapproaching seawater compositionin a hydrothermal fault zone merits discussion.
The Late Tortonian marine flooding of the fault zone best explains this counterintuitive isotopic shift. Seawater percolated downward into the fault zone where it was heated, and later re-ascended. The volume of seawater relative to hydrothermal fluids in the shallow domain of a submerged fault zone is high enough that a shift towards seawater chemistry is reasonable (Fig. 15). Comparing reconstructed d 13 C DIC values of Late Triassic seawater (brachiopod shells; Saltzman & Thomas, 2012) to Middle Miocene to Recent oceans, an obvious overlap is observed (Fig. 13). The implication of this is a convoluted differentiation between 'typical' Late Triassic and 'typical' Middle Miocene seawater values. Furthermore, d 13 C values compiled here plot within the range of either of these potential source waters. Consideration regarding mineralogy-dependent (calcite versus dolomite) fractionation are moot, because the enrichment of 13 C in dolomite relative to calcite (ca 1&; Swart, 2015) is only significant in fluids at Earth surface conditions, and is insignificant at elevated fluid temperatures such as in burial environments. In principle, two end-member interpretations exist. The first being that dolomites preserved a primarily Late Triassic seawater d 13 C DIC signature that was not significantly altered during subsequent burial and overprint in the fault zone. This interpretation assumes a rock-controlled diagenetic environment with regard to d 13 C. The absence of 13 C-enrichment from the degassing of magmatic CO 2 from intrusions into the fault zone provides circumstantial evidence of this. Carbon isotope fractionation during the uptake of magmatically-derived CO 2 from HCO 3-in marine waters (Chacko et al., 2001, and references therein) yields approximately +3&, which is believed to sufficiently enrich the local marine d 13 C signature (Jacquemyn et al., 2014) Alternatively, Fig. 13 illustrates that all dolomite d 13 C values plot in the lower portions (ca 0 to +2&) of the reconstructed Triassic seawater values that overlap with the range of Miocene to Recent seawater d 13 C DIC . This interpretation suggests a pervasive re-equilibration of all dolomite phases with respect to carbon isotopes, and a shift towards Miocene seawater d 13 C DIC . This controversy cannot presently be resolved by means of carbon isotope data alone, and additional lines of evidence must be considered, including 87 Sr/ 86 Sr ratios (Fig. 16) and petrography. The observation of well-preserved zoning of Dol 3 to Dol 8 cements ( Fig. 6E and H) contrasts the theory of pervasive carbon isotope re-equilibration with Miocene seawater. Although some of these zones are internally patchy ( Fig. 6F and H), zoning at a scale of some tens of micrometres is well-preserved. Given that pervasive re-equilibration of dolomite d 13 C values is only possible via a microscale dissolution-reprecipitation mechanism, and would be expected to obliterate the subtle chemical zonation, the possibility of this mechanism can be excluded.
All studied dolomite phases plot outside of reconstructed Middle/Late Triassic seawater 87 Sr/ 86 Sr ratios and shift towards the possibly more radiogenic Miocene ratios (Fig. 16). As expected, the cataclastic matrix represents an endmember stage of dolomite alteration and hydrothermal overprint, with 87 Sr/ 86 Sr C1 ratios in reasonable agreement with Miocene data, or are even more radiogenic. The latter observation is best explained by the highly radiogenic values of crust-derived fluids that interacted with Rbbearing minerals, typical for deep-reaching fault zones ( Fig. 16; Faure et al., 1963;Perry & Turekian, 1974;Dong et al., 2017). However, hydrothermal fluids from magmatic bodies in (or near) the Carboneras Fault Zone must also be considered (Benito et al., 1999;Mattei et al.,   2014) as sources for highly-radiogenic strontium. In contrast to the cataclasite values, the early diagenetic dolomicrite phases (Dol 1 and Dol 2) plot near Triassic seawater values. However, even small amounts of material with a highly radiogenic signature of crust-derived fluids would affect bulk samples to a disproportional degree, such as the case for Dol 1. This is exemplified in values of autochthonous Neogene calcitic units that plot significantly outside of the range of Miocene seawater (Fig. 16). Assessing all available data, a significant reequilibration of Triassic 87 Sr/ 86 Sr dolomite ratios and carbon isotope values in the presence of Miocene (or younger) seawater or crust-derived hydrothermal fluids (Fig. 15) seems unlikely, due to: (i) lack of evidence for a 13 C-enriched CO 2 source for d 13 C dolomite ; (ii) a moderate shift of bulk dolostones towards more radiogenic 87 Sr/ 86 Sr ratios; (iii) the contamination bias of Dol 1 samples with material yielding radiogenic Sr; and (iv) a well-preserved (albeit patchy) zonation ( Fig. 6F and H) of dolomite cements, ruling out bulk dissolution-reprecipitation.

Marine transgressive deposits and late stage meteoric diagenesis
The isotope geochemistry of marine bivalve shells sampled in the autochthonous Azagador Formation that transgressed the fault zone displays typical Miocene marine calcite values. In contrast, bulk sediment samples are significantly more negative in both their oxygen and carbon composition. Connate fluids present during deposition of these and other Neogene strata from the Carboneras Nijar Basin had no observable impact on impermeable Alpujarride dolostones. The latest paragenetic stage, characterized by a regression and subaerial exposure of the sea floor, is identified by authigenic quartz (Qz 1), and blocky low-Mg and high-Mg calcite meteoric cements (LMC 1 to LMC 4 and HMC 1 to HMC 3; Fig. 11). In the field area, dolostones are typically covered by a decimetre-thick calcitic caliche soil, i.e. a subrecent weathering product of carbonates typical for a semi-arid climate (James & Choquette, 1990). Carbon isotope values of late meteoric-vadose calcite cements, speleothems and caliche are between À10& and À5&. Studies have proven a significant impact on d 13 C compositions (À10 to 2&) of Caribbean Pleistocene LSTs driven by 12 C contribution associated with the decay of organic material (Lohmann, 1987;James & Choquette, 1990). Values for the aforementioned meteoric-vadose samples reported here (calcite cements, speleothems and caliche) agree with these findings. However, a significant overprint of paragenetically older dolomite cement phases was not found. This may be related to low fluid circulation within the low porosity and permeability dolostone, and high resistance of these phases to changes in their d 13 C composition at near-surface conditions. In summary, the geologicallyyounger regressive intervals and the subsequent meteoric stages did not significantly affect the dolostones of interest here.
The multi-proxy approach conducted here includes radiometric age dating of specific cement phases and allows for a detailed reconstruction of a complex history of dolomitization and tectonic/ hydrothermal overprint. Refer to Fig. 17 for an overview of all depositional, diagenetic and tectonic stages including the reconstructed burial history of the Alpujarride dolostones.  (Lonergan & Platt, 1995;Vissers, 2012;Rutter et al., 2012, and references therein). Johnson (1993) proposed peak temperatures of 200°C based on fission track analysis, whereas Lonergan & Platt (1995) proposed that the stratigraphically lowermost Malaguide dolostones reached low-grade anchizone conditions. Contrastingly, Critelli et al. (2008) analyzed red beds interstratified with Malaguide dolostones and proposed burial temperatures of 100 to 160°C and lithostatic/tectonic loading of 4 km. Independent of these varying interpretations, the differences in burial depth and temperatures between Alpujarride and Malaguide strata resulted in two different diagenetic pathways, as supported by the contrasting data found here (Bakker et al., 1989;Vissers, 2012). It is plausible that a detailed comparison of these two dolostone units could have its merits, but a different study area that offers equal access to both units must be chosen.  Fig. 17. Interpreted burial history, tectonic events and the associated diagenetic processes affecting Alpujarride dolostones from deposition in the Middle/Late Triassic to Recent. Large-scale tectonism includes the rotation of Iberia, Alpine Orogeny and Carboneras Fault Zone activity. Interpreted maximum burial depth of ca 10 to 15 km for Alpujarride dolostones is given in green, whereas bulk Alpujarride rocks (25 km) after Bakker et al. (1989) and Vissers (2012) are shown in orange. Associated depths during precipitation correspond to pressure correction of fluid inclusion analysis. Burial depth not at uniform scale.

Implications for ancient dolostone archives
In the following, data shown here are critically set in context with previously published work utilizing marine or early diagenetic dolostones as palaeoenvironmental archives. The number of papers dealing with dolomite as palaeoenvironmental archives for the Phanerozoic is comparably limited (e.g. Richter et al., 2014;Geske et al., 2015b;Hood et al., 2018, and references therein), and is clearly related to the relative scarcity of Phanerozoic early diagenetic dolomite compared to the abundant biogenic material and bulk carbonate micrite (Turpin et al., 2014, and references therein). In the absence of biogenic carbonates in the deep time rock record, dolomite archives become disproportionally more important (Tucker, 1982;Burns & Matter, 1993;Halverson et al., 2005;Mazumdar & Strauss, 2006;Kasemann et al., 2014;Cui et al., 2017; and references therein). Complications arise from the fact that Precambrian dolomite archives are either direct marine precipitates (Tucker, 1983;Hood et al., 2011;Wood et al., 2017;Shuster et al., 2018), early to late diagenetic replacement phases of aragonitic, calcitic or dolomitic precursors, or direct pore-filling precipitates of deep-seated hydrothermal fluids. Work shown here and in previous studies, represents a clear note of caution that the uncritical use of geochemical data from bulk dolostones, possibly mixed with that from LSTs, must be avoided (e.g. Magaritz et al., 1986;Lambert et al., 1987;Jenkins et al., 1992;Shen et al., 1998;Melezhik et al., 2009). More recently, the number of studies critically balancing environmental and diagenetic patterns is increasing (e.g. Kaufman & Knoll, 1995;Knauth & Kennedy, 2009;Derry, 2010a;Grotzinger et al., 2011). Well-constrained studies disentangling Neoproterozoic cement parageneses include Cui et al. (2016cCui et al. ( , 2017, Zhou et al. (2016) or Hood et al. (2018). Moreover, non-conventional proxies, such as boron isotopes and calcium isotopes (Kasemann et al., 2005), magnesium isotopes (Kasemann et al., 2014), clumped isotope palaeothermometry (Loyd et al., 2015) and cadmium isotopes (Hohl et al., 2017), may offer a new perspective of the Neoproterozoic world. Using the Carboneras Fault Zone dolostones to test the resistance of these proxies to complex diagenetic overprint may be worthwhile.
Regarding the Neoproterozoic rock record, many dolomitization models, as based on Phanerozoic case studies, must be considered with caution due to some fundamentally different environmental parameters (Algeo et al., 2015;Canfield et al., 2008). An issue with the interpretation of Ediacarian rocks is the uncertainty of seawater d 18 O values. Many models exist, but none are based on empirical evidence, or able to answer all associated questions (see Cui et al., 2017, and references therein). Taking the results of this study and previous work into account, it is clear that d 18 O dolomite signatures are generally prone to diagenesis, and may even react to retrograde alteration. In the case example shown here, a D 18 O of 4& between the early marine dolomicrite Dol 1 and cataclasite C 1 is observed. The authors acknowledge that most would not consider Proterozoic cataclasite as archives of marine isotope values, but the general warning is clear, that even in diagenetically-stable carbonate phases such as stoichiometric dolomite, d 18 O values are potentially compromised. In contrast, dolomite d 13 C signatures in rock-buffered, fluidlean systems are remarkably well-preserved even under extreme conditions. Despite all of these issues, it is encouraging that early diagenetic (fabric-preserving) replacement dolomites precipitated from (near) marine porewaters, are possible archives for marine climate dynamics of the ancient Earth, independent of the chosen time interval. Problems that may result when exploiting early diagenetic dolomites lie in their often non-stoichiometric, Ca-rich nature (Geske et al., 2015b) and from any complex amorphous precursors (Mavromatis et al., 2017). When the subsequent stabilization to a stoichiometric Mg/CaCO 3 occurs in the presence of marine shallow burial fluids, a diagenetically-stable archive is formed. This study provides evidence for this concept and encourages the use of a broad suite of analytical proxies. For example, and as suggested by Derry (2010a), fluid inclusion studies of Neoproterozoic rocks are still surprisingly scarce. Data reported here, and in many previous studies, demonstrate that fluid inclusion analysis is a powerful tool to examine primary fluid inclusion temperature ranges from 50 to 250°C, even in dolostones that have experienced low-grade overprint at temperatures up to 300°C.

CONCLUSIONS
The multi-proxy study discussed here documents a test of preservation potential of marine dolomite archive proxy data to extreme, longterm diagenesis, including low-grade metamorphosis. As the product of these events, eight dolomite phases, a dolomitic cataclasite and a series of meteoric/vadose calcite cements resulted, and are here characterized in terms of petrographic, paragenetic and geochemical criteria.
The complex geological history of these dolostones commenced in the Middle to Late Triassic with protolith deposition and fabricpreserving, early marine to shallow marine burial replacive dolomitization (Dol 1) at maximum fluid temperatures of 50°C. During intermediate burial, zebra saddle dolomite formed while the bulk rock further recrystallized at temperatures up to 150°C (Dol 2). Phase Dol 1 recrystallization and Dol 2 formation likely occurred during an interval of large-scale tectonism related to the rotation of Iberia and the opening of the Proto-Atlantic Ocean during the Early-Middle Jurassic, when deep hydrothermal fluids ascended to shallower levels. Corrosion of Dol 3 cement marks a change to undersaturation in the fluids with respect to dolomite. Phases Dol 4 and Dol 5 are likely also related to the rotation of Iberia, and have peak fluid inclusion homogenization temperatures of 250°C. Dolomite phases 1 to 5 were precipitated in a compactional regime that did not significantly change over ca 100 Myr. Phase Dol 6 represents a relatively quiescent duration in terms of precipitation, as a consequence of low remaining porosity due to physical and chemical compaction. The subsequent Dol 7 and Dol 8 cement phases signify a fundamental change to a deeper tectonic setting during an early phase of the Alpine Orogeny. Highgrade anchizone conditions did not cause dolostones to evolve into marble fabrics, and these rocks have striking textural preservation as documented by their primary fluid inclusions. Timing of deep, hot, dolomite cement precipitation and subsequent homogenization of d 13 C and d 18 O occurred during the Early Eocene to Early Miocene. During the main phase of the fault zone activity (Middle to Late Serravallian), the dolostones experienced an increasing degree of brecciation and cataclasation, coupled with a significant fluid-dominated shift in dolomite d 18 O values. Despite these multi-stage tectonohydrothermal events, early marine diagenetic replacive dolomites preserved their Late Triassic seawater dissolved inorganic carbon (DIC) and 87 Sr/ 86 Sr signatures. This is considered to be evidence for an uncommonly conservative nature of some, but not all, of these geochemical proxies.
The term 'dolomite' should be used with caution, because it is sometimes used in a simplified manner for an often complex suite of early marine to late burial Mg-carbonates, each of which represent different physico-chemical parameters and proxy data. Therefore, the interpretation of dolostone data without a robust petrographic framework is discouraged. One should also acknowledge that isotope or elemental proxy data will likely vary in their recalibration or stabilization based on a fluid-buffered or rock-buffered system. This implies that the same dolomite archive may yield well-preserved or highly-altered data depending on the analyzed proxy and diagenetic environment. Evidence is given that petrographic criteria for alteration does not necessarily imply either a resetting or preservation of the proxy data.

ACKNOWLEDGEMENTS
This study was performed in the context of the collaborative research initiative CHARON Phase II (DFG Forschergruppe 1644). The authors thank B. Gehnen and K. Krimmler from the isotope geochemistry laboratory for isotope analysis, S. Weisel, M. Born, A. Schulz and T. Seemann from the thin section preparation department, T. Reinecke and H. Mammen for X-ray diffraction analysis, and N. J€ ons for microprobe analysis at Ruhr-University Bochum. Additional thanks go to R. Coimbra from the University of Aveiro for help with regional geological literature and to R. Stritzke from the Geological Survey of North Rhine-Westphalia for spores and pollen analysis in dolostones. Special thanks go to R. Hoffmann, A. Hueter, L. Wulff and Y. Hu for their time and helpful comments which helped to improve this paper, and to J. Parr for assistance in sample preparation and drawing of figures. The authors express their gratitude to the journal editors Dr G. Della Porta and Dr H. Qing, as well as to reviewers Dr D. Morrow and an anonymous colleague. Special thanks go to Dr Morrow for his extremely thorough review. Sampling of rocks was not possible in the Cabo de Gata Natural Parc due to legal restrictions. However, the Carboneras Fault Zone expands beyond the boundaries of the Natural Park, where sampling of all relevant lithologies is possible. The authors also made use of extensive rock collections at the Ruhr-University Bochum that stem from several unpublished PhD projects performed in the 1970s and 1980s in this region.