δ60Ni and δ13C Composition of Serpentinites and Carbonates of the Tekirova Ophiolite, Turkey, and Meatiq Ophiolite, Egypt

Nickel isotope fractionation patterns in continental ultramafic environments generally show a depletion of δ60Ni in weathered rocks and an enrichment in bedrock samples. The present study focuses on stable Ni isotope fractionation patterns in carbonate-rich, ultramafic ophiolite samples with concomitant fluids at an active serpentinization site in southwestern Turkey, with a comparison to results from an inactive serpentinization site in the Eastern Desert of Egypt with carbonate-rich samples. All solid phase data from the inactive serpentinization area are consistent with previously reported values from serpentinites, whereas the solid precipitates in the active area (SW Turkey) give values slightly heavier than previously reported data. However, the Ni isotopic signatures in the active serpentinization system likely reflect the scavenging of light Ni by iron oxide and carbonate precipitation, as has been previously demonstrated in laboratory coprecipitation experiments. It is also possible that the active system results resemble previous laboratory experimental results that show a relatively strong initial fractionation between fluids and solids, which then diminishes with time due to aging of the precipitates.


INTRODUCTION
Several studies have investigated stable Ni isotope compositions in ultramafic/mafic rocks to better understand the geochemical cycling of Ni in natural environments (Cameron et al., 2009;Gueguen et al., 2013;Ratié et al., 2015;Spivak-Birndorf et al., 2018). It has been shown that the world's major rivers are enriched in heavier Ni isotopes (+0.29 to +1.34‰) compared with solid Ni sources (+0.15‰), which, in contrast, are depleted in 60 Ni (Cameron and Vance, 2014). The Ni isotopic composition of unweathered mafic and ultramafic rocks (such as gabbro, basalt, peridotite, dunite, and serpentinized varieties) falls within a relatively narrow range of −0.13-+0.32‰ (Gall et al., 2013;Gueguen et al., 2013;Ratié et al., 2015;Spivak-Birndorf et al., 2018). Weathering of mafic and ultramafic rocks results in extensive formation of clays and iron oxides, which have both been shown to preferentially adsorb light Ni (Wasylenki et al., 2015;Spivak-Birndorf et al., 2018). In contrast, Nirich hydrous Mg-silicates, such as garnierite, a common product of serpentinization, show enrichment in heavy Ni isotopes (Spivak-Birndorf et al., 2018). No directed analyses were performed on other common natural serpentinization products such as brucite (Mg-hydroxide) or Mg-Ca carbonates. However, experimental adsorption and coprecipitation of calcite and Ni have shown a preferential adsorption/incorporation of light Ni into calcite (Alvarez, 2019) with a Δ 60 Ni fluid−solid as low as −1‰. Strong hydration of the divalent Ni ion at high calcite precipitation rates is one explanation for this behavior. It is still unknown how aging of calcite influences the isotope signature of Ni isotopes in experimentally precipitated calcites. In a coprecipitation study on Mn-oxides and Ni by Wasylenki et al. (2019), it was shown that the initial, strongly fractionated Δ 60 Ni fluid−solid signature diminished after a few weeks due to aging. This effect would potentially obstruct the possibility of finding natural counterparts unless samples were analyzed immediately after collection. Aging is a physical transformation process in which the precipitated minerals may change mineralogy, morphology and composition with time. Element substitution and ratios may change significantly with time, as shown by Wasylenki et al. (2019), where the concentration and thus the isotopic signature of Ni in the mineral precipitate changes.
Terrestrial areas of active serpentinization are commonly associated with low vegetation, high-pH fluids, high dissolved Ca and Mg concentrations, reduced gas seepage, and extensive secondary mineral precipitation (Etiope et al., 2011). This study focuses on the Tekirova ophiolite (near Antalya, Turkey), where localized and diffusive fluxes of abiotically formed CH 4 gas seep from the rock as a product of active low-temperature serpentinization (Etiope et al., 2011;Etiope and Sherwood Lollar, 2013). High pH and extensive carbonate precipitation are common features of the area. As a comparison, a serpentinized carbonate-rich ophiolite with no active serpentinization in the Meatiq area, Eastern Desert (ED), Egypt, is investigated. In order to better understand the geochemical cycling of Ni in carbonate-rich ultramafic and mafic rocks, this study aims to investigate the C and Ni isotopic differences between active and inactive serpentinization systems with extensive precipitation to evaluate the influence of secondary mineral precipitation on the fractionation of stable Ni isotopes. Carbon isotopes were additionally measured to evaluate the possible influence of microorganisms on the Ni isotopic fractionation. The two field sites of the current study contain fresh and aged carbonate precipitates as well as serpentinites, brucites, and iron oxides and may therefore display a wide variety of δ 60 Ni signatures, from very depleted, associated with iron oxides, to enriched, commonly associated with the meteoric waters flowing on the ophiolite surface (Cameron and Vance, 2014). Both study sites have been or are currently subjected to carbonation, a process in which igneous rocks interact with fluids rich in CO 2 , resulting in carbonate precipitation.

Geological Setting of the Tekirova Ophiolite Outcrop
The studied site is a terrestrial methane gas seep at the Tekirova ophiolite in Turkey. The Tekirova ophiolite is a part of several N-S trending zones of rocks associated with rifting, basin formation, and development of passive margins (Hosgörmez, 2007). Tekirova is located in the tectonic contact zone between the western Kemer Zone serpentines and the Tekirova Zone carbonates, all forming part of the larger Antalya Complex, which was emplaced onto the continental plate in the Late Cretaceous. During the emplacement, ophiolitic material was incorporated into the continental carbonates, initiating a strong serpentinization process and secondary carbonate precipitation. Well-preserved ultramafic to mafic cumulate rocks crop out at Çıralı, where serpentinization-induced gas seeps are located (Hosgörmez, 2007). The western Kemer Zone consists of a Mesozoic carbonate platform that overlies Ordovician-Triassic sedimentary rocks. The eastern Tekirova Zone, on the Mediterranean coast, is composed of Cretaceous ophiolite with well-preserved olivine-rich ultramafic to mafic cumulates as well as peridotite, gabbro, serpentinized harzburgite with minor lherzolite, podiform dunite, and chromitite (Aldanmaz et al., 2009). Between the two zones lies the Tekirova fault, where serpentines and metasediments are the dominant rocks. Below the fault are several rock formations that can generate hydrocarbons: the Sapandere bituminous shale (Ordovician-Silurian), the Beydagları limestone (Lower Mesozoic), and the Pamucakyayla siltstones and coal (Carboniferous). An extensional phase started in the late Pliocene and is ongoing, which may form pathways for gas seepage due to the resulting increase in the permeability of the rocks (Demirel and Gunay, 2000). Active serpentinization is associated with the ophiolitic outcrop, forming extensive seeps of methane gas continuously flowing from the rocks, visible as continuously burning flames. Active serpentinization is also evident in the mineralogy of the area, which shows serpentine clasts in host minerals, indicative of serpentinization reactions at greater depths involving temperatures between 200 and 300°C (Hosgörmez et al., 2008;Etiope et al., 2011;Etiope and Sherwood Lollar, 2013). Serpentine, brucite, and magnetite were produced during hydration reactions of olivine and pyroxene. The mineral textures also show that further serpentinization reactions within the ophiolite proceed at low temperatures (below 200°C) and are currently active, forming the secondary carbonate veins (Zwicker et al., 2018). Earlier studies have reported that the fluids in the area have pH values of 7-12 and gas compositions of 87% CH 4 , 7.5-11% H 2 , 2-4.9% N 2 , 0.57% light alkanes, 0.01-0.07% CO 2 , and 80 ppm He (Woodcock and Robertson, 1977;Meyer-Dombard et al., 2015. The sources of these gases are both abiotic and biotic, as shown by stable isotopic analyses (Hosgörmez et al., 2008). The most common minerals in the rocks at Tekirova are, in decreasing order of bulk abundance, serpentine (lizardite-chrysotile), brucite, calcite, aragonite, magnesian calcite, chromite, rutile, and ferromanganese (Hosgörmez, 2007).
(Atg-) rich serpentinites, as well as metagabbro and metabasalt and their metamorphosed and carbonate-altered products such as talc-rich rocks, talc-carbonate rocks, and listvenites. The degree of carbonate alteration is more intensive in Atgserpentinites than in Lz-serpentinites. The ophiolitic rocks from the Meatiq area are included in this study as a reference  representing ophiolitic rocks with no active serpentinization. The ophiolite emplacement occurred after approximately 736 Ma. Alteration of the ophiolite, such as serpentinization and carbonation due to interaction with migrating carbonate-rich fluids (Stern and Gwinn, 1990), occurred at ∼600 Ma (Boskabadi et al., 2017).

Sampling
A total of eight solid samples and nine water samples were collected at Tekirova in Turkey, and 12 solid samples were collected at Meatiq in Egypt ( Figure 1; Table 1). Sediment samples were collected using a sterile spatula and stored in sterile Falcon tubes.

Tekirova
The sampling sites are situated at an exposed area located on the Tekirova ophiolite outcrop, north of Çıralı, Antalya, Turkey. Flames of burning gas are scattered around the site, and three small fluid seeps meander downslope over the outcrop. The site is mostly dry during the year, but during the rainy season (November-February), small streams and ponds of standing water are present not only on the primary sampling site but also close to other sampling sites located at higher elevations. Onsite pH measurements were obtained using a hand-held pH meter (pHTestr 30, Thermo Scientific, Eutech Instruments Pte. Ltd., Germany, with an uncertainty of approximately 0.2 pH units).
The Tekirova samples were collected in February, during the rainy season, which was necessary in order to be able to collect any fluids since the site is often completely dry. Fluid samples for trace element analyses were collected from sites JSB, SB, LB1, LB2, FG1, and FG2 (Table 1) by withdrawing fluids using sterile syringes and were injected into acid-washed (3 M HCl for 24 h) 10 ml plastic tubes and sealed. After collection, all samples (except samples for Ni isotopes, which instead were centrifuged) were filtered through a pre-cleaned cellulose nitrate 0.2 μm filter into a pre-cleaned plastic tube and stored at +8°C. Sediment samples (FG1, FG2, and IR) were collected at the same time as the water samples by scooping the material into sterile plastic tubes using pre-cleaned plastic spoons. Solid samples for mineral analyses were collected using a rock hammer and plastic bags and were collected at the same sites as the fluid samples.

Meatiq
The Meatiq area is located about 40 km west of the Red Sea, Egypt. Ophiolitic rocks of the Meatiq area comprise mainly serpentinites and their altered products such as talc-rich rocks, talc-carbonates, and listvenites. U-Pb dating of zircons from an ophiolitic gabbro (Fawakhir, located SW of the Meatiq area) suggests a crystallization age of 736.5 ± 1.2 Ma for the ophiolites in this region (Andresen et al., 2009). A detailed description of the geology of the area can be found in Boskabadi et al. (2017). The samples from the Meatiq area were collected at three main localities from quarry and/or roadside outcrops and were selected based on the type of serpentinization, the extent of carbonation, and the carbonated products. A detailed description of the sampling in the Meatiq area can be found in Boskabadi et al. (2017).

Analysis of Elements in Solution
The dissolved elemental compositions of the fluid samples were measured using inductively coupled plasma-optical emission spectroscopy (ICP-OES, Spectro, Varian Vista AX), with Ar as the carrier gas, at the Department of Geological Sciences, Stockholm University, Sweden. Approximately 10 ml of filtered fluid sample were mixed with 0.5 μl > 65% HNO 3 (Sigma-Aldrich, puriss.) prior to analysis. The analytical error was ∼4%. All measurements were obtained using multi-element matrix-matched standards and certified standards (NIST 1640a) as a control for the multi-element standards (LGC-Promochem) used for inductively coupled plasma mass spectroscopy (ICP-MS).

Optical Microscopy
Optical microscopy of thin sections was conducted using a Leitz polarization microscope to determine mineralogy. An Olympus BX-51 fluorescence microscope equipped with an Olympus DP-71 digital camera was used to determine the biologic influence and/ or presence in the samples (excitation wavelength of 395-440 nm, emission >470 nm), where a green color represents carbonate precipitates (dark green) and a yellow color (dark yellow) represents organic matter (Dupraz et al., 2013).

Raman Spectroscopy
Raman spectra were collected using a confocal laser Raman microspectrometer Horiba instrument (LabRAM HR 800; Horiba Jobin Yvon) equipped with a multichannel air-cooled (−70°C) 1,024 × 256 pixel charge-coupled device (CCD) detector. Acquisitions were obtained with an 1,800 lines/mm grating. Excitation was provided by an Ar-ion laser source (λ 514 nm). Spectra were recorded using a laser power of 0.1-1 mW at the sample surface. A small quantity of the sample material was placed on a glass slide on an Olympus BX41 microscope coupled to the instrument. The laser beam was focused through a 100× objective to obtain a spot size of about 1 μm. Collection times for the Raman spectra were 10 accumulations of 10 s each. The accuracy of the instrument was controlled by repeated use of a silicon (Si) reference with a Raman line at 520.7 cm −1 . The Raman spectra were obtained with LabSpec five software.

Scanning Electron Microscopy
Surface composition and topography were analyzed semiquantitatively on an XL30 environmental scanning electron microscope (ESEM) with a field emission gun (FEG). The ESEM was equipped with an Oxford X-Act energy dispersive spectrometer (EDS), a backscatter electron detector (BSE), and a secondary electron detector (SE). Peak and elemental analyses were performed using INCA Suite 4.11 software. None of the samples were coated.

Carbon Analysis (Dissolved Inorganic Carbon, Dissolved Organic Carbon, and Stable Isotopes)
Stable carbon analyses were performed to investigate the influence of microorganisms on the secondary mineralization Frontiers in Earth Science | www.frontiersin.org June 2021 | Volume 9 | Article 651967 and the fluids. Isotope analysis of dissolved inorganic carbon (DIC) were made by direct injection of 5 ml of water sample into 12 ml septum-seal glass vials (Labco Limited, cat. No. 238 W) containing 100 μl of 85% H 3 PO 4 , and flushing with He gas (100 ml/min) for 5 min. The method used for dissolved organic carbon (DOC) is described in detail in Lang et al. (2012). A volume of 4 ml of a sample or standard was transferred into an exetainer vial using a disposable borosilicate pipette, and 1 ml of the oxidizing agent was then added. The samples were acidified to pH < 3 with 8.5% H 3 PO 4 (85% Merck Analysis grade orthophosphoric acid; Merck, Darmstadt, Germany) and diluted by a factor of ten with Milli-Q water (Millipore, Billerica, MA, United States). To remove inorganic carbon, high-purity helium was bubbled (60 ml/min) through the sample manually for 10 min by piercing the septum with a flushing needle. The septum was also pierced with a second needle for the output gas stream. This outlet needle was connected to a copper tube whose open end was submerged in water to prevent any backflow of atmospheric CO 2 into the sample. The DIC and DOC analyses were performed at the Department of Environmental Science and Analytical Chemistry (ACES), Stockholm University, using a Gasbench II coupled to a MAT 253 mass spectrometer, both from Thermo Scientific. Measurement reproducibility was calculated to be better than 0.1‰ for δ 13 C. Solid sample analyses were performed using an Elemental Analyzer NC 2500 (Carlo Erba) coupled via a split interface to a mass spectrometer (DeltaV Advantage, Thermo Scientific). Precision and accuracy were monitored by reference to long-term analyses of laboratory and international standards. Results are reported as δ‰ values relative to the VPDB scale for δ 13 C. Carbon isotope analyses of the Meatiq samples were carried out on the carbonate (mainly magnesite) fraction of serpentinite whole-rock samples (see Boskabadi et al., 2017, for more details).

Ni Isotopic Reagents and Material
Nickel isotopic analysis was performed in a clean room (Class 1000, or ISO-6) at the Ifremer Centre de Bretagne, Brest, France, following the method described by Gueguen et al. (2013) and using a double spike method. Samples and standards (NIST SRM 986) containing Ni concentrations of 0.5-1 μg Ni were evaporated on a hotplate. The spike (containing equal proportions of 61 Ni and 62 Ni) was then added to all samples and standards at a Ni sample /Ni spike ratio of one and re-evaporated to dryness.

Sample Preparation and Rock Digestion
Rock and water samples were digested in a mixture of HCl-HNO 3 in polytetrafluoroethylene vials and evaporated until dry, followed by addition of HF (28 mol L −1 trace metal grade or Suprapure grade) to dissolve all solid remnants. The samples were then evaporated to dryness at a temperature of 70-80°C. The HF treatment step was repeated to achieve complete digestion of nonsiliceous rocks and ensure removal of potential fluorides formed in the course of the first dissolution step for siliceous materials. Ni concentrations were re-determined before Ni elution using an ICP-Quadrupole (ICP-Q-MS, X-series II) at Pôle Spectrométrie Océan, Brest. Aliquots (<1 ml) were withdrawn from the water samples and rock solutions and diluted with 0.28 M HNO 3 to a final volume of 10 ml, and Ni concentrations were then measured. With known element concentrations, an optimal concentration of the double spike could be added to trace possible interfering elements.

Sample Purification
Dry sample-spike residues were dissolved in 1 ml 0.24 M HCl. Thereafter, 0.3 ml 1 M ammonium citrate and 125 μl concentrated ammonia were added to increase pH. Approximately 0.5 ml wet Ni-spec resin (Ni Eichrome) was loaded into polypropylene columns. The resin was then washed with 6.5 ml Milli-Q water and 4.5 ml of 0.2 M ammonium citrate +0.45 M ammonia and capped. An additional 0.5 ml of the 0.2 M ammonium citrate +0.45 M ammonia solution was added to buffer the resin pH to ∼9. The prepared sample solution was then loaded onto columns containing dimethylglyoxime (DMG) molecules, which at pH < 8 form a Ni-DMG complex that is retained in the resin. The resin was resuspended, increasing the active surface area and rate of complexation, and left for 1 h. The Ni yields through the column were close to 100%, implying that the 1 h equilibration time was sufficient. The caps were removed, and the columns were allowed to drain. After draining, 2.2 ml of the 0.2 M ammonium citrate +0.45 M ammonia solution was added twice, at intervals of 15 min, to wash out Fe and other matrix elements. 2.2 ml of Milli-Q water was added three times at intervals of 15 min to wash out excess ammonium citrate and ammonia salts. Ni was then collected in 6.6 ml of 3 M HNO 3 (added in three applications of 2.2 ml at 15 min intervals). The collected Ni fractions were then dried down on a hot plate at ∼95°C overnight. The dry sample residue was then dissolved in 1 ml of 0.3 M HNO 3 for isotopic analysis.

Mass Spectrometry
Ni isotopic compositions were determined using a Thermo Scientific ™ Neptune multicollector inductively coupled plasma mass spectrometer (MC-ICP-MS) at the Ifremer Centre de Bretagne in Brest, France. Samples and standards were introduced via an ApexQ MicroFlow PFA-50 self-aspirating nebulizer at an injection rate of ∼60 µl min −1 . The sample and skimmer cones were made of Ni, but 58 Ni intensities of the blanks indicated that no Ni was contributed from the cones, as demonstrated previously (Moynier et al., 2007;Fujii et al., 2011;Gueguen et al., 2013;Neubeck et al., 2021).
Each set of samples was run twice. A short preliminary run (10 cycles, each lasting 4 s) was performed to better determine Ni, Fe, and Cu concentrations. For every sample, there was a standard with a similar concentration and Ni spike /Ni natural ratio. Samples were generally grouped in sets of 2-5 samples, preceded and succeeded by one bracketing standard of NIST SRM 986. In the second run, performed to obtain high-precision isotope data, each sample was measured through five blocks of 10 cycles with 4 s of integration time apiece, separated by an instrument rinse of 10 cycles with 4 s of integration time. The instrument can very precisely measure samples with concentrations over 0.03 ppm Ni, with 2SE < 0.1. Lower concentration measurements yield slightly larger uncertainties in δ-values, with 2SE ≈ 0.1‰. Only concentrations under 0.01 ppm have extremely high uncertainties exceeding 1‰ (Gueguen et al., 2013). Ni concentrations in rock samples were very high (>10 ppm), and the Ni concentrations in water samples were generally sufficient for precise measurement.

Ni Isotope Data Reduction
Nickel isotope compositions of samples are expressed using the delta (δ) notation, in which the isotopic ratio of the sample is normalized to that of the NIST SRM 986 bracketing standard, following the equation: where X denotes either the 60 Ni, 61 Ni, or 62 Ni isotope ( 64 Ni was not considered in this study). The data presented in this study is described using the 60 Ni/ 58 Ni ratio given as δ 60 Ni. The Ni isotopic compositions of samples were then calculated using the double spike calculation template from Gueguen et al. (2013). The NIST SRM 986 standard was passed through columns four separate times, yielding an average value and external reproducibility for the method of −0.01 ± 0.08‰. Procedural blanks yielded an average amount of Ni of approximately 4 ng.

Tekirova-Mineralogy
Tekirova host rock is composed of strongly serpentinized peridotite with extensive precipitation of iron oxides and carbonates (Hosgörmez et al., 2008;Meyer-Dombard et al., 2015). According to optical microscopy analyses and SEM, the LB samples were dominated by serpentine minerals with minor carbonate precipitates (mainly calcite and minor amounts of aragonite, Figure 2; Supplementary Figure S1). The BFE sample was dominated by brucite and hydromagnesite with minor amounts of carbonates (Supplementary Figure S2), whereas the BF1 and BF2 samples were almost completely dominated by carbonates (Supplementary Figure S3). The IR, FG1, and FG2 samples consisted of loose sediments; IR and FG2 consisted of iron oxides and carbonates (Supplementary Figures  S4, S5), while FG1 consisted purely of carbonates (Supplementary Figure S6). The BF samples are dominated by concentric layers of calcite. Fine and coarse-grained layers of carbonate occur with sharp contacts from yellow to dark yellow/brown, fibrous, and mesh-textured chrysotile.
Micro-stromatolites and microspheres (Ø 20-25 µm) with layers of calcite alternating with organic material are abundant in the samples, and green fluorescence was used to distinguish between the calcitic and organic material (Figure 3). Distinct differences between organic and abiotic (chemically precipitated) layers could be observed by Raman spectroscopy and fluorescence microscopy.
Sample IR is completely covered by microorganisms, and organic compounds, cells, and filaments are visible in the SEM images and cover the entire sediment surface (Figure 4). The Raman spectra of the IR sample show carbonaceous matter and iron oxides.

Major and Minor Elemental Composition of Tekirova Fluids
The LB samples show a slight downhill increase in pH from 7.6 to 8.0, whereas the FG samples show the opposite trend of a downhill decrease in pH from 10.7 to 8.7. Both BF1 and BF2 show a very high  Frontiers in Earth Science | www.frontiersin.org June 2021 | Volume 9 | Article 651967 7 pH of 11.4. The IR sample has a pH of 7.9, and the JSB and SB river samples give pH values slightly above neutral, of 8.0-9.5, respectively. The color changes from white to red in the FG1 and FG2 samples is coupled with a strong increase in Fe from 5.6 to 83.5 ppm and a decrease in Ca from 73.1 to 41.1 ppm. The concentration of Ni in the fluids decreases downhill from 4.5 to 0.8 ppm at LB1 and LB2, respectively, whereas the opposite can be observed in FG1 and FG2, where the Ni concentration increases from 0.3 to 4.5 ppm. The rest of the samples show relatively similar Ni concentrations of around 0.4 ppm, except BF2, which has a slightly higher Ni concentration of 1.1 ppm.

Meatiq Mineralogy
Three types of rock samples were collected in the Meatiq ophiolite in Egypt ( Figure 5): 1) Lz-serpentinite composed mainly of lizardite and later chrysotile veinlets with minor carbonate, traces of chlorite, and bastite pseudomorphs; 2) Atgserpentinite consisting mainly of antigorite and carbonate with minor talc and/or clinochlore; and 3) talc-rich rock, composed mainly of talc with minor amphibole, clinochlore, and/or quartzcarbonate veins, and that has undergone extensive mylonitization. The common features of all the Meatiq samples are the lack of active serpentinization and the presence of carbonate alteration that increases from Lzserpentinite to Atg-serpentinite and talc-rich rocks. The sampling site is completely dry, and thus no fluid samples were collected (for field site images, see Boskabadi et al., 2017). δ 60 Ni and δ 13 C Composition of Fluid and Solid Samples at Tekirova The dissolved Ni in fluids shows strong variations in both total Ni concentration and δ 60 Ni values: from 0.21 to 0.96 ppm and +0.41 ± 0.09 to +1.77 ± 0.26‰, respectively. Solid δ 60 Ni values are dominantly positive, ranging from −0.05 ± 0.07‰ to +0.38 ± 0.06‰ ( Figure 6; Table 2) with an average value of +0.19 ± 0.26‰ (n 6), all of which lie within previously reported ranges for mafic and ultramafic rocks (Ratié et al., 2015). The serpentine-rich, dark-colored BFE samples have δ 60 Ni values (+0.25 ± 0.03‰) similar to those of the light-colored calcites, brucite-rich veins, and fracture fillings, with δ 60 Ni values of +0.18 ± 0.06‰. Corresponding δ 13 C values are −20.53 ± 0.15‰ for serpentines and heavier values of −15.56 ± 0.15‰ for the brucites and veins. Calcite-dominated samples show δ 13 C values of approximately −23‰. The IR sample shows the lightest values of both δ 60 Ni and δ 13 C, −0.05 ± 0.07‰ and −30.19 ± 0.15‰, respectively. The Lz-and Atg-serpentinites and the talc-rich rocks from the Meatiq area show homogeneous δ 60 Ni values ranging from +0.09 ± 0.04‰ to +0.16 ± 0.05‰, in agreement with earlier reported values for serpentinites (Gall et al., 2013;Cameron and Vance, 2014;Ratié et al., 2015).

DISCUSSION
The main difference between Tekirova and Meatiq is the degree of serpentinization and the activity at the two sites. Meatiq, situated in the Egyptian desert, is an inactive system without any circulating fluids or gas seeps characteristic of active serpentinization, such as fluids with high pH and seepages of reduced gases such as CH 4 and/or H 2 (Sánchez-Murillo et al., 2014). Tekirova, by contrast, has both alkaline fluids with elevated pH and gases characteristic of active serpentinization (CH 4 , H 2 , and CO 2 ). Both sites are strongly serpentinized. Meatiq is almost completely serpentinized, with no preserved olivines or pyroxenes (Boskabadi et al., 2017), whereas the Tekirova ophiolite shows a degree of serpentinization between 30 and 65% (Etiope et al., 2011). Both ophiolites show extensive carbonate precipitation, with the distinction that carbonate precipitation at Tekirova is an ongoing event, while only aged carbonates are found at Meatiq.

Fluids at Tekirova
As previously described in Meyer-Dombard et al. (2015), the Tekirova waters are of meteoric origin, affected by active serpentinization at the site. The fluids can be divided into two representative serpentinization water types: 1) high pH and high Ca, and 2) neutral pH and high Ca (Figure 3). Compared with similar serpentine seeps, our data represents either a water type more similar to some waters in Cyprus (Neal and Shand, 2002) and Taro-Ceno (Boschetti and Toscani, 2008) with high Ca and neutral to slightly above neutral pH, or a fluid type more similar to Ca-OH − type waters found in Tekirova (Meyer-Dombard et al., 2015), Voltri (Cipolli et al., 2004), New Caledonia, Oman, and Yugoslavia (Barnes et al., 1978). In contrast to the extremely high concentration of Ca in fluid data presented by Meyer-Dombard et al. (2015), our Ca concentrations are similar to those reported from other serpentine systems ( Figure 6; Barnes et al., 1978;Cipolli et al., 2004;Boschetti and Toscani, 2008). The contrasting concentrations of Ca may be due to natural fluctuations in the serpentinization rate as well as the overall fluid flow in the area. Both studies were conducted in February during the rainy season, but in contrast to the single fluid seep sampled by Meyer-Dombard et al., we discovered three seeps at the site ( Figure 6). Thus, a higher fluid flow may explain our diluted values in sampled waters compared to the study by Meyer-Dombard et al. (2015). The intermittent nature of the spring waters of Tekirova, resulting in local and temporal compositional differences, may also explain differences in elemental composition. The larger rivers, SB and JSB, have considerably lower Ni concentrations than the meandering streams on the ophiolite outcrop, with average concentrations of 5.3 and 6.1 nM, respectively, but showed Ni concentrations similar to previously reported values in the Kalix River in Arctic Sweden (2-6 nM) and parts of the Amazon River (4-5 nM; Cameron and Vance, 2014). The meandering streams on the ophiolite, however, show values similar to previously reported values from the Ottawa  Gueguen et al., 2013;Ratié et al., 2015;Steele et al., 2011;Gleeson et al., 2004). The two water types are shown in the ovals, where A represents low-pH waters and B high-pH waters.
The fluid pathways at Tekirova are dominated by percolation of rainwater through the permeable limestone down to the groundwater (Neal and Shand, 2002). Progressive supersaturation of rainwater percolating through the limestone and the contact with high-pH serpentinization fluids initiates carbonate precipitation (visible as red or white precipitates ( Figure 6) and a concomitant drop in pH. This drop in pH was measured at FG1 and FG2, to pH values of 10.7 and 8.7, respectively. A downhill decrease in the Ca/K ratio from 110 to 57.3 (Table 3) is also found to accompany extensive, clearly visible carbonate precipitation.

Carbon Isotopes
Carbon isotopic data of the fluids and solids are very similar to the reported values by Meyer-Dombard et al. (2015). The organic carbon influence decreases down-channel, which is evident also in our data where the total δ 13 C is −26.14‰ at FG1 compared with −18.98‰ at FG2 downhill. At FG2, the DIC δ 13 C values (δDI 13 C) are all close to atmospheric values ( Table 2) and are thus interpreted as dominantly chemically precipitated calcite, whereas the measured DOC δ 13 C (δDO 13 C) isotope ratios are indicative of a mixed influence of abiotic seep gases (with a measured δ 13 C range between −18 and −26‰; Etiope et al., 2011;Etiope and Sherwood Lollar, 2013) and thermogenic organic matter, as previously reported by Etiope et al. (2011) and Etiope and Sherwood Lollar (2013). At the IR site, however, the δ 13 C is as low as −30.2‰, which is likely due in part to the extensive microbial colonization shown in Figure 4. The emanating gases were previously discussed by Etiope and Sherwood Lollar (2013), who concluded that the majority of the gas (80-90%) is of abiotic origin, that the remaining 10-20% is thermogenic or microbial, and that the measured δ 13 C of carbonates (approximately −11‰; Etiope and Sherwood Lollar, 2013) closely reflects the gas composition. Our δ 13 C of −30.2‰ at IR, which is slightly more negative than the previously reported values of −26‰ from the area, may therefore reflect a larger influence of the microbial colonization visible in the sample (Figure 7). The large amounts of iron oxides in the IR sample are likely not accumulating 12 C (see Zhang et al., 2021) and are thus not likely to be the major cause of the slightly more negative signal seen in the samples. Zhang et al. instead observed an enrichment of 13 C in iron oxides, which would, if this process were also prominent at IR, indicate an even stronger fractionation by the microorganisms in the IR sample.

Ni Isotopes
Meteoric fluids generally show a negative correlation between δ 60 Ni and [Ni] (Cameron and Vance, 2014), and this is also true for the Tekirova LB samples (Figure 7;   Table 2), similar to what has been reported by Cameron and Vance (2014) and Cameron et al. (2009) for secondary mineralization. The two FG sites differ in mineral composition; FG2 is composed of iron oxides and carbonates, whereas FG1 is composed purely of carbonates ( Table 2). It has been shown that light Ni isotopes are preferentially incorporated into iron oxides and calcites during coprecipitation experiments (Eusterhues et al., 2011;Wasylenki et al., 2015;Alvarez, 2019;Neubeck et al., 2021). The preferential uptake of the lighter isotope is coupled to a large surface area and/or growth rate (Alvarez, 2019;Hiemstra et al., 2019). Calcite growth rate is controlled by surface complexation that in turn is controlled by supersaturation and pH (Ruiz-Agudo et al., 2011). At pH values below 8.5, calcite grows mainly by the incorporation of CaCO 0 3 at the surface, whereas at pH values greater than 9, Ca 2+ is also a contributor to the crystal growth due to a faster water exchange rate in the CaCO 0 3 ligand. A faster water exchange rate leads to slower calcite precipitation growth, allowing for Ni, which has a slow water exchange rate, to be incorporated into the  (2014), marked with an asterisk. Error bars are for long-term reproducibility of standards analyzed using the double spike method. The horizontal gray band gives the average and 1SD of Ni isotope compositions of samples of continental silicate rocks (Cameron et al., 2009 precipitating calcite. Since light Ni is preferentially incorporated into the solid phase, fractionation of Ni isotopes will thus increase with a decreased calcite precipitation rate at higher pH. This effect is likely not very strong at the LB site, where the pH increase is minor. More interesting, then, is the opposite trend observed at the FG sites, where a strong Ni isotope fractionation from +0.43 to +1.2‰ coincides with a dramatic pH drop from 10.7 to 8.7. If calcite precipitation were the only controlling factor for Ni isotopic fractionation, the opposite trend would be observed here. Therefore, we suggest that the precipitation of iron oxides rather than calcite at FG2 controls the precipitation and fractionation of Ni isotopes. This is supported by the downhill mineralogy that shows clear evidence of iron oxides in the solids both in the Raman spectroscopy analyses and in the change of color from white precipitates at FG1 to rusty red at FG2 ( Figure 6). The formation of iron oxides is also strongly influenced by pH (Zhu et al., 2016). Any iron-rich fluid that comes in contact with a high-pH fluid such as the FG fluids would immediately induce iron oxide precipitation. This extensive iron oxide precipitation is evident at FG2 both in the Raman spectra ( Figure 4) that coincide with Raman spectra of natural hematite (De Faria and Lopez, 2007) and in the strong color of the precipitates (Figure 7). The δ 60 Ni signature of +1.2‰ in the fluids at FG2 is consistent with the documented strong fractionation effect that iron oxide precipitates have on Ni isotopes (Wasylenki et al., 2015;Neubeck et al., 2021). At site FG2, the Δ 60 Ni solid−fluid is as large as −1.01‰, comparable to reported values of −0.77‰ for adsorption of Ni to goethite under experimental conditions (Gueguen et al., 2013). In Alvarez (2019), carbonate coprecipitation experiments reported Δ 60 Ni solid−fluid values close to −1‰, which indicates a stronger effect than our reported value at FG1, where a majority of the precipitates are carbonates, but the Δ 60 Ni solid−fluid values are only −0.03‰. This indicates that calcite precipitation has a minor effect on δ 60 Ni at FG1.
All solid samples (except the IR site with a dominant iron oxide composition and a δ 60 Ni of −0.05 ± 0.07‰, see below) had δ 60 Ni values ranging from +0.18 ± 0.02‰ to +0.38 ± 0.06‰, significantly higher than previously reported δ 60 Ni data from olivines/peridotites (−0.26-0.12‰; Gueguen et al., 2013;Cameron and Vance, 2014;Figure 8) or serpentines (−0.13 ± 0.04‰; Gueguen et al., 2013) but similar to reported values for saprolite, which is rich in clays and goethite (Ratié et al., 2015). Meyer-Dombard et al. (2015) showed the presence of clay minerals (montmorillonite) in some samples analyzed from the Tekirova ophiolite, which, if also present in our samples, could likely have had an effect on the δ 60 Ni values. No XRD analyses were performed in this study, so the influence of clays on the δ 60 Ni values at Tekirova has yet to be analyzed. The δ 60 Ni values of the two rivers, SB (+0.90 ± 0.2‰) and JSB (+1.06 ± 0.07‰), are similar to values measured in other rivers globally, with values ranging from +0.31‰ (Rio Negro; Cameron and Vance, 2014) to +1.34‰ (Yangtze; Cameron and Vance, 2014). Even though these values fall within the global range, they are still quite high, likely due to the meandering through areas of active serpentinization, extensive precipitation of light Ni into iron oxides, and possibly other secondary mineralization, leaving the rivers enriched in δ 60 Ni.
Our enriched δ 60 Ni data are thus likely due to the ongoing serpentinization with extensive precipitation. The aggregated results from our study indicate that most of our solid samples show positive δ 60 Ni values. Interestingly, the δ 60 Ni values of the solid BF samples were similar to the reported values for soil samples (Ratié et al., 2015) or mafic/ultramafic rocks (Gueguen  (Ratié et al., 2015) but depleted compared with the other local Tekirova samples. This signature coincides with the most negative value of δ 13 C ( Table 2). A possible explanation could be the influence of microorganisms and/or iron oxides at the site (Neubeck et al., 2017). Microorganisms, especially methanogens, have been experimentally shown to strongly fractionate stable Ni isotopes (Cameron et al., 2009). Since many methanogens are hydrogenotrophs, using H 2 and CO 2 as energy and carbon sources, the documented seeps of both CO 2 and H 2 (Etiope et al., 2011) at the Tekirova active serpentinization system are likely contributors to fractionation of Ni isotopes. In a recent paper by Zwicker et al. (2018), lipid biomarker evidence for methanogens and methanotrophs was detected within the rocks in the Tekirova ophiolite, strengthening the inference of microbial influence on the IR sample. The SEM images ( Figure 2) show a clear difference between IR and the other samples, in that the IR sample is completely covered with microorganisms and iron oxides (Figure 3). Iron oxides have been experimentally shown to adsorb the lighter Ni isotope and may therefore have an impact on the measured values at the IR site. Other influencing factors on the depleted Ni value are coprecipitation, biologic uptake, or ligand binding (Wasylenki et al., 2007;Fujii et al., 2011;Wasylenki et al., 2015). IR is the only site where light C and Ni coincide, indicating that light Ni may not be easily preserved unless under extensive influence from microorganisms.

Meatiq Ni Isotopes
The solid δ 60 Ni values of the Meatiq serpentinite all fall within previously reported values for peridotite (Gueguen et al., 2013) and Bulk Silicate Earth (Steele et al., 2011;Figure 8). No statistical difference can be seen in δ 60 Ni average values between the Atgserpentinite, Lz-serpentinite and talc-rich samples, with average values of 0.14 ± 0.05, 0.12 ± 0.1, and 0.12 ± 0.09‰, respectively. However, plotting δ 60 Ni as a function of [Ni] shows a positive correlation between δ 60 Ni and [Ni] in the Atg and Lz samples, whereas no correlation can be observed in the talc-rich samples ( Figure 9). The talc-rich samples represent greenschist facies high in carbonate (dolomites and magnesites) and low in sulfides, similar to the Atg samples that are subjected to high-temperature alteration (350-400°C), which are rich in carbonates and low in sulfides. The Lz sample, on the other hand, contains few carbonates and high amounts of sulfides (pentlandite and millerite), and represents lower alteration temperatures FIGURE 9 | Meatiq sample δ 60 Ni plotted as a function of Ni concentration (Barnes et al., 1978;Neal and Shand, 2002;Cipolli et al., 2004;Boschetti andToscani, 2008, Morrill et al., 2013;Meyer-Dombard et al., 2015).