A deep fluid source of radiogenic Sr and highly dynamic seepage conditions recorded in Miocene seep carbonates of the northern Apennines (Italy)

The 87 Sr/ 86 Sr, REE and Mo-U systematics were investigated in Miocene seep carbonates formed on accretionary ridges at the front of the northern Apennine wedge. Here we evaluate fluid sources, seepage intensity and redox conditions during carbonate precipitation. Micrite matrix and early calcite cements lining cavities are the main authigenic carbonate phases and show negligible diagenetic alteration. Their AOM-related (anaerobic oxidation of methane) origin is evident from  13 C values ranging from -38.7‰ to -25.7‰. The 87 Sr/ 86 Sr ratios in carbonates vary between 0.708659 and 0.709132. Most micrite values fall within the range of Middle Miocene seawater (Langhian, MNN5a biozone), in agreement with the biostratigraphy of the host sediments, thus reflecting precipitation of seep carbonates close to the seafloor. Highly radiogenic 87 Sr/ 86 Sr ratios, of early calcite cements and micrite from conduit-rich facies at the base of carbonate bodies, likely derived from the interaction of fluids with detrital clays during the fault-controlled upward migration through the underlying terrigenous turbidite


Introduction
Marine seepage accounts today for ~20 Tg per year of methane emissions (Anderson et al., 2012).An evaluation of its contribution to global greenhouse gas emissions requires a deep understanding of the process of anaerobic oxidation of methane, and its spatial and temporal variability (AOM, Boetius et al., 2000).Our knowledge of cold seeps has greatly benefited from the study of methane-derived carbonates from the fossil record, which reflect the long-term evolution of seepage systems in space and time (e.g.Campbell, 2006;Conti et al., 2010Conti et al., , 2017;;Feng et al., 2009a, b;Jakubowicz et al., 2015;Malone et al., 2012).Geochemical data on seep carbonates informs on fluid expulsion processes, composition of precipitating fluids, seepage intensity and rates of AOM (e.g.Aloisi et al., 2000;Greinert et al., 2001;Malone et al., 2012;Naehr et al., 2007;Peckmann et al., 2001;Roberts, 2001).Moreover, the application of Strontium Isotope Stratigraphy (SIS, McArthur et al., 2001) to authigenic carbonates has proved to be a successful tool to obtain reliable formation ages (Ge and Jiang, 2013;Kiel et al., 2014;Kiel and Hansen, 2015).In fact, the 87 Sr/ 86 Sr isotope composition of seep carbonates reflects the Sr signature of the precipitating fluids (Burke et al., 1982;Joseph et al., 2012;Solomon et al., 2009;Teichert et al., 2005;Torres et al., 2004;Veizer, 1989) and seep carbonates forming close to the seafloor generally show 87 Sr/ 86 Sr values of coeval seawater (Naehr et al., 2000).Deviation from the expected Sr isotope ratios in weakly altered carbonates possibly indicates mixing with fluids modified by fluid/rock interaction processes (Hong et al., 2018;Joseph et al., 2012;Sample et al., 1993;Tong et al., 2013;Viola et al., 2017).Such a complex situation is common at accretionary wedges.In these settings, carbonates precipitate from a mix of seawater and deeper fluids, bringing hydrocarbons toward the seafloor through different migration pathways (Joseph et al. 2012;Teichert et al. 2005;Torres et al. 2004).
In authigenic carbonates, rare earth elements (REE) are commonly used to reconstruct interstitial redox conditions and to trace the evolution of early diagenetic fluids (e.g.Birgel et al.,

A C C E P T E D M
A N U S C R I P T 2011; Himmler et al., 2010;Hu et al., 2014;McRae et al., 1992, Wang et al., 2015;Zwicker et al., 2018).Carbonates forming under normal marine conditions in oxygenated waters follow the smooth seawater REE profile, characterized by enrichments in HREE along with positive La and negative Ce anomalies (Elderfield et al., 1990).Deviations from the seawater pattern may correspond to differential incorporation of redox sensitive elements (such as Ce, MREE) under suboxic or anoxic conditions, or be related to admixture with fluids contaminated by fluid-rock interaction processes (Bayon et al., 2011;Frimmel, 2010;Kim et al., 2012;Sarkar et al., 2003).
In oxic pore-waters, Ce(III) is partially oxidized to Ce(IV) on the surface of Fe-Mn oxyhydroxides, leaving a typical negative Ce anomaly in fluids, which is eventually recorded by authigenic phases within the sediments (Frimmel, 2010;Shields and Webb, 2004).Conversely, positive Ce anomalies represent precipitation in anoxic to suboxic porewaters.Methane-derived carbonates are characterized by a wide range of Ce concentrations, with positive anomalies indicative of precipitation under strictly anoxic conditions and negative or absent anomalies ascribed to episodic oxic conditions and seawater influences (Feng et al., 2010;Hu et al., 2014;Wang et al., 2015).
The recent application of redox sensitive Mo and U to the study of sediments and authigenic carbonates has advanced the comprehension of seepage dynamics and identification of paleo-seeps within the sedimentary column (Chen et al., 2016;Hu et al., 2017).In oxygenated seawater, Mo is mainly present as molybdate anion (MoO 4 -2 ) at uniformly high concentrations (~105 nM ;Collier, 1985).Molybdenum is relatively unreactive in oxic conditions and is removed by slow adsorption to Mn-Fe oxyhydroxides (Morford et al., 2009;Zheng et al., 2000).
Uranium concentration in carbonates provides further insights into porewater redox conditions.Under oxic conditions, dissolved U occurs as U(VI) forming uranyl and uranyl

A C C E P T E D M
A N U S C R I P T carbonate ions and shows a conservative geochemical behavior (Algeo and Tribovillard, 2009;Cumberland et al., 2016;Russell et al., 1994).Sediment uptake of U starts in suboxic porewaters at the Fe(III)-Fe(II) transition, when the soluble U(VI) is reduced to insoluble U(IV).Due to the larger size of uranyl ion relative to Ca(II), its partition coefficient in carbonate is low.The incorporation of U into authigenic carbonate phases is favored by a much higher partition coefficient in suboxic and anoxic porewaters because the ionic radius of U(IV) is similar to that of Ca(II) (Sturchio et al., 1998).The degree of enrichment of Mo and U relative to the average continental crust (Taylor and McLennan, 1985) allows a better understanding of the variations in porewater conditions and AOM intensity through time (Algeo and Tribovillard, 2009;Hu et al., 2014).Since U enrichment occurs within the Fe reduction zone, (Mo/U) EF < 3.1 (Mo/U of present-day seawatwer) is generally associated with suboxic to anoxic conditions.Sulfidic interstitial conditions promote Mo accumulation and progressively increase the (Mo/U) EF ratio of sediments up to (Mo/U) EF >3.1 (Algeo and Tribovillard, 2009).
In this paper, we report the 87 Sr/ 86 Sr ratios, REE and Mo-U systematics of Miocene seep carbonates formed atop an intrabasinal high in the inner foredeep of the Italian northern Apennines (Argentino et al., 2019).Active seeps from thrust-related ancticlines have been the object of several studies due to their close association with shallow gas hydrates (e.g.Greinert et al., 2001;Teichert et al., 2005;Torres et al., 2004); however, fossil examples are still poorly reported in the literature (Campbell, 2006 and references therein) or the structural setting often lack sufficient constraints for an accurate comparison with modern counterparts.This study aims to identify the origin of fluids involved in authigenic carbonate formation at a Miocene seep, to assess the contribution from deeply-derived fluids and to reconstruct seepage dynamics.We also tested the Strontium Isotope Stratigraphy (SIS) methodology (McArthur et al., 2001) on seeprelated micrites using nannofossil biostratigraphy of the host sediments to corroborate the results.

The northern Apennines
ACCEPTED MANUSCRIPT The northern Apennine chain is an orogenic NE-verging wedge, characterized by the stacking of several structural units of oceanic and continental origin (Fig. 1).The complex structure of the chain is the result of the convergence and collision between the European and Africa plates, with the interposition of Adria and Corsica-Sardinia microplates, from the Mesozoic to the present (Carminati and Doglioni, 2012).The convergence stage started at the end of the Early Cretaceous, caused by the progressive consumption of the Piedmont-Ligurian Ocean, a portion of the Tethys.The complete closure of the ocean during Middle-Late Eocene caused the rapid uplift and erosion of the Alpine orogenic wedge and the inception of the continental collision.During the early stage of the collision (late Oligocene), the internal oceanic units (Ligurian nappe), which were deformed and accreted during the eo-meso-Alpine phases, were thrusted over the adjacent thinned margin of the continental Adria microplate, represented by Subligurian units.From Miocene to Recent, the thrust system migrated towards the foreland, involving the Tuscan and Umbrian-Marchean units (Fig. 1) deposited on the Adria microplate.
The collisional stage was marked by the flexuring of the foreland and the formation of foredeep basins, progressively migrating towards NE (Argnani and Ricci Lucchi, 2001;Tinterri and Magalhaes, 2011).Small basins developed on top of the migrating accretionary wedge, filled by Epiligurian units (Conti et al., 2016;Ricci Lucchi, 1986).Thrust-related anticlines, occurring as intrabasinal highs in the foredeep, formed from the Burdigalian to the Messinian during different migratory phases of the foreland system (Conti et al., 2010, 2017 andreferences therein).
Sedimentation on top of the intrabasinal highs (Fig. 1C) was mainly represented by hemipelagites and diluted turbidites (drape mudstones) up to hundred meter thick fine-grained intervals.Large seep carbonate bodies are hosted within these pelitic intervals at the front of the orogenic wedge and testify for a link between the seepage distribution and tectonic activity (Argentino et al., 2019).

The Corella outcrop
The Corella outcrop is located in the Tuscan Apennines, ~25 km northeast of Florence ([43°56'59"N 11°34'59.8"E]Fig. 1A).Seep carbonates are enclosed in a pelitic interval striking NW-SE, parallel to the structural trends of the chain and in proximity of the thrusting of the Tuscan units over the Marnoso-arenacea Fm.This interval is ~12 km long and up to 200 m in

ACCEPTED MANUSCRIPT
A C C E P T E D M A N U S C R I P T thickess (Conti et al., 2017).The outcrop consists of six large authigenic bodies distributed in two stratigraphic horizons (L1, L3, L5 in the lower, and L2, L4, L6 in the upper; Fig. 1B).
Carbonate bodies show stratiform to lenticular morphologies, up to 250 m in width and maximum thickness of 35 m, and are mainly formed of authigenic micrite locally enclosing finegrained siliciclastic components.The basal portion of the carbonate bodies is dominated by conduit-rich facies and breccias with millimetre-sized clasts.Chemosynthetic communities, represented by articulated lucinids and vesicomyids, have been observed during fieldwork (Argentino et al., 2019).Based on calcareous nannofossil biostratigraphy of the enclosing marls, this outcrop has been ascribed to the Langhian MNN5a subzone (Conti et al., 2017).

Methods
Twenty-one samples of authigenic carbonates were collected from the Corella outcrop (Fig. 1).Most of the specimens derive from the carbonate body L1, which is the best exposed and easy to access; other samples were collected from L2, L4 and L5 carbonate bodies (Fig. 1B).We have identified three different carbonate components based on petrographic observations (Fig. 2A,B): 1) authigenic micrite, dark grey to brown in color, locally rich in pyrite and with dispersed fine siliciclastic grains (matrix micrite); 2) early calcite cement lining and filling cavities and fractures (rim cement), commonly formed by anhedral to subhedral cloudy crystals; 3) lucinid shell fragments.The center of cavities and veins are filled by large, generally limpid, blocky calcite crystals likely related to late diagenetic processes that have not been considered in this study (Fig. 2).Table 1 reports sample location and the different types of carbonate components.
Carbon and oxygen stable isotopes were analyzed in 10 specimens of matrix micrite and rim cement obtained from 5 samples.Powders were obtained from polished slabs using a handheld microdrill and about 10 mg were reacted with 100% phosphoric acid at 25 °C for 24 h.et al., 2004) was analyzed as an unknown and its repeated measurements were used to check the analytical precision and accuracy.Average values of the selected masses in JCt-1 agree within the uncertainty of the recommended values (Okai et al., 2004).The 44 Ca concentration in each sample was determined by ICP-OES and used as an internal standard.We employed NIST 610, NIST 612 and NIST 614 (Jochum et al., 2011;Pearce et al., 1997)  For Ca, Al, Sr and Mn determinations, 5 mg of powders were dissolved in 1 ml 16 N HNO 3 and 1 ml concentrated HF in Teflon beakers.The sealed beakers were left on a hot plate at 90 °C for ~24 h.Solutions were then dried and re-dissolved in 4% HNO 3 .Analyses were performed on a Perkin-Elmer Optima 4200DV ICP-OES hosted at the Department of Chemical and Geological Sciences, University of Modena and Reggio Emilia.Precision RSD (n=3) was better than 5% for Ca and better than 2% for Al, Sr and Mn.
For Sr isotope analysis, about 10 mg of powder was obtained by microdrilling polished surfaces of the rock slabs, subsequently rinsed three times with MilliQ water.After drying, each sample was leached two times using a solution of 0.3% w/w acetic acid to sequentially dissolve about 30% and 40% of the powdered material, following the procedure described in Li et al.
Sr was separated from the matrix by ion exchange chromatography following procedures reported in Argentino et al. (2017a) and Weber et al. (2018).The whole procedure was conducted in a clean laboratory under a laminar flow hood, with a typical Sr blank <100 pg.
Strontium isotope ratios were measured by the MC-ICP-MS (Thermo Scientific Neptune) housed at the CIGS (University of Modena and Reggio Emilia).Solutions were diluted to ~50 ppb and introduced into the Neptune using a desolvating system (ESI-APEX IR) and a 100 μl/min nebulizer.Samples and standards were analyzed in a static multicollection mode in a single block of 100 cycles with an integration time of 8.4 s per cycle.Mass bias normalization was performed using the exponential law and a 87 Sr/ 86 Sr ratio of 8.37520.Isobaric interferences of 86 Kr and 87 Rb on 86 Sr and 87 Sr species were corrected.The 87 Sr/ 86 Sr ratio was corrected for instrumental bias to a NBS-987 value of 0.710248 (Thirlwall, 1991).Repeated measurements of the NBS-987 standard during two analytical sessions yielded a mean 87 Sr/ 86 Sr value of 0.710233 ± 0.000014 and 0.710228 ± 0.000012 (2σ, n=19 in both sessions). 87Sr/ 86 Sr ratios were converted into numerical ages using the regression curves LOWESS look-up Table version 5 (McArthur et al., 2012).Maximum and minimum ages were obtained respectively subtracting and adding the cumulative error (reported in parenthesis) to the mean isotopic value.
In this paper, we calculated REE anomalies as follow: Ce/Ce* = 3Ce N /(2La N + Nd N ), where N refers to concentration normalized to the Post Archean Australian Shale (PAAS; Taylor and McLennan, 1985).Enrichment factors for Mo, U and As (respectively Mo EF , U EF and As EF ) were defined as follows: X EF = [(X / Al) sample /(X / Al) reference], where X and Al are the weight concentrations of elements X and Al, respectively (Tribovillard et al., 2006).X EF >3 represents a detectable authigenic enrichment whereas X EF >10 is indicative of a moderate to strong enrichment (Algeo and Tribovillard, 2009).

A C C E P T E D M
A N U S C R I P T

Stable C and O isotopes
The stable carbon and oxygen isotopes composition of the methane-derived carbonates

Sr isotopes
The 87 Sr/ 86 Sr ratios in micrite from the carbonate body L1 range from 0.708659 to 0.709132 (Fig. 3).Fifteen 87 Sr/ 86 Sr values from L1 are concentrated in a narrow range, with values comprised between 0.708713 and 0.708808 (Tab.1), whereas one sample (3M) at the base of the body, is remarkably more radiogenic with 87 Sr/ 86 Sr ratio of 0.709132.Four micritic samples collected from the basal portion of the other carbonate bodies (L2, L4, L5) are characterized by more radiogenic Sr isotope ratios with values between 0.708837 and 0.708980.
Rim cements in all examined bodies (L1, L2, L5) show isotopic ratios from 0.708761 to 0.709132, generally higher than micrite.A lucinid shell collected at the base of L2 gives a 87 Sr/ 86 Sr ratio of 0.708802 (Tab.1).In general, the most radiogenic specimens (3V and 3M, L4, L2_VA and L2_VB) are from the conduit-rich facies in the basal portion of bodies.
Cerium anomalies range between 0.7 and 1 while La and Eu anomalies are comprised between 1.3 and 6.4, and between 1.1 and 2.6, respectively.
(Mo/U) EF ratios show a broad range of values from 0.

Assessment of diagenetic alteration
Ancient marine carbonates may retain the original 87 Sr/ 86 Sr composition when unaffected by diagenetic alteration.This is also true if they have interacted with fluids characterized by a similar isotopic signature during early diagenesis or with fluids with low Sr concentration (Denison et al 1994).Post-depositional processes within carbonates generally lead to a decrease in Sr, Na, Mg and increase in Mn, Fe and Zn, since the concentrations of these elements are different in seawater and diagenetic fluids (Brand and Veizer, 1980).Manganese and Fe 2+ show a similar behavior in solution and a comparable distribution coefficient in carbonates (D>1); they are more soluble in reducing fluids and substitute for Ca 2+ in the carbonate lattice.Strontium tends to concentrate in interstitial fluids (D<1) (Tucker and Wright, 1990).Thus, ancient seep carbonates preserving the original Sr isotopic composition are commonly characterized by Mn content <300 mg/kg (Ge and Jiang, 2013;Kiel et al., 2014).In our study, Mn values are below this threshold (Tab. 1) and consistent with literature reference values for well-preserved seep carbonates (Joseph et al., 2013) and normal marine limestones (e.g. Brand and Veizer, 1980;Denison et al. 1994).Moreover, Mn and Sr do not correlate with each other (Fig. 4A) or with 87 Sr/ 86 Sr ratios.Only rim cements 87 Sr/ 86 Sr ratios show a positive correlation with Sr (R 2 =0.87) (Fig. 4B).However, this behavior is opposed to what we would expect for meteoric alteration and may reflect the original mineralogy, namely higher Sr in calcite recrystallized from aragonite.Carbonates with  18 O signature close to coeval seawater are more likely to have preserved the original isotopic signal (Kiel et al., 2014).Our micrite  18 O composition ranges from -2.8 ‰ to 0.3 ‰, close to the Miocene seawater signature (Tab.2) (Zachos et al., 2001).
The oxygen isotope composition of rim cements is more depleted than the adjacent micrite, but their  13 C is similar and their Mn and Sr concentrations suggest a limited diagenetic alteration.It is possible that other factors such as fluid temperature may have controlled the oxygen isotope signature of rim cements.Additional evidences for negligible diagenetic alteration are the poor correlation between  18 O and  13 C (R 2 = 0.28, Fig. 4C), and the absence of mixing lines of  13 C and  18 O with 87 Sr/ 86 Sr ratios (R 2 = 0.14 for  13 C, R 2 = 0.35 for  18 O, Fig. 4D).Based on these observations, we assume that our samples preserved the original Sr isotope composition.

Sr isotope dating and fluid source characterization
Most samples of micrite in this study have 87 Sr/ 86 Sr ratios corresponding to Langhian seawater Sr composition (MNN5a biozone) (Tab.1; Fig. 5), in agreement with the biostratigraphic age of the host sediments (Conti et al., 2017).These data indicate that carbonate precipitation occurred in equilibrium with coeval seawater or shallow porewaters (Peckmann et al., 2001) Sample 15M Sr isotope composition corresponds to a converted age of 16.9 Ma, which is older than the host sediments.Although the analysis was conducted on a rather homogeneous micritic sample, (assessed by petrographic observations), some contamination from fine-grained detrital carbonate particles may have contributed to the older age (Fig. 5).
Recently, rim cements of seep carbonates have been identified as the preferable SIS target given that they are very dense and less susceptible to interaction with late diagenetic fluids (Kiel et al., 2014).However, they are sometimes very thin and difficult to sample and contamination with adjacent material cannot be completely ruled out.Our data show that Sr analysis of the micritic component gives reliable results and may extend the field of applicability of SIS to seep carbonates.However, a detailed assessment of the diagenetic alteration is always recommended prior to isotopic analyses.
Early calcite cements lining cavities and fractures located in the basal portion of the bodies (Fig. 3), are characterized by more radiogenic 87 Sr/ 86 Sr ratios, which in some samples (3M, 3V, L2_VB, L4, 2VA, 2VB) are remarkably high, with values up to 0.709132.These radiogenic ratios cannot be explained in terms of absolute age, but they reflect deviation from coeval seawater and, therefore, the involvement of fluids of different origin.It is widely accepted that alteration of clay minerals, deriving from weathering of continental crust with high Rb/Sr ratios, progressively releases radiogenic Sr into interstitial waters, thus accounting for the precipitation of carbonate cements with high Sr isotope composition (Sample et al., 1993;Veizer et al., 1989).In our case study, the underlying thick terrigenous turbidites of the Marnoso arenacea Fm. and deeper pelitic units rich in clay minerals, might have contributed the radiogenic Sr.Some minor contributions from detrital micas might be also expected (e.g.Peckmann et al., 2001).Diagenetic fluids enriched with radiogenic Sr were then conveyed toward to the seafloor by a thrust-fault active during the Middle Miocene at the front of the Apenninic wedge (Fig. 1C).A similar interpretation has also been proposed for radiogenic 87 Sr/ 86 Sr values in carbonates from a vertical fault at the deformation front of the Oregon accretionary wedge (Sample et al., 1993) and for the Shenhu seep carbonates (continental slope,

A C C E P T E D M
A N U S C R I P T South China Sea; Tong et al., 2013).It is worth mentioning that locally, during carbonate precipitation, the deep radiogenic Sr signal can be masked by mixing with seawater, as the amount of Sr in seawater is much higher than in the seeping fluid (Tran et al., 2014).

Rare Earth Elements
Diagenetic processes generally alter the carbonate REE composition producing a negative correlation between Ce/Ce* and Dy N ⁄Sm N , and a positive correlation between Ce/Ce* and ∑REE (Shields and Stille, 2001).No strong correlations are evident in our samples, so we can assume that they retained the original REE budget (Fig. 8A) (Shields and Stille, 2001) They define a diagenetic trend in terms of Pr/Yb ratio and Mg/Sr ratio.Our data fall outside this trend, having higher Pr/Yb ratios and more variable Mg/Sr ratio.In addition, the Mn and Fe contents are low.Therefore, the REE abundances of calcites have not been affected by laterstage porefluid diagenesis and reflect the original composition of early phases.This is also based on the assessment of diagenetic alteration discussed in section 5.1, and on the relationships between Ce/Ce* and Dy N /Sm N (Fig. 8A) and between Ce/Ce* and ∑REE (Fig. 8B).
The ∑REE of the micrites ranges between 8.6 and 17.2 mg/kg (Tab.3), similarly to modern seep

A C C E P T E D M A N U S C R I P T
carbonates from the Gulf of Mexico (Feng et al., 2009b;Hu et al., 2014), Congo deep-sea fan (Feng et al., 2010), South China Sea (Ge et al., 2010), Gulf of Cadiz (Wang et al., 2015) as well as some ancient seep carbonates (Feng et al., 2009a).Rim cements tend to have lower REE, from 0.1 to 14.9 mg/kg, similar to values reported in Himmler et al., (2010) for microcrystalline and botryoidal aragonite.Variations in the REE of seep carbonates are linked to variations in the carbonate precipitation rate and seepage intensity.Slow rates favor the incorporation of REE and lead to high ∑REE concentrations (Ge et al., 2010), similarly to authigenic apatite (Wright et al., 1987).Variability in the REE content of authigenic carbonates can also be related to mixing with seawater, given the low ∑REE of the latter (Himmler et al., 2010).It is therefore difficult to prove whether the broad range of REE values reported in the present study is due to variable precipitation rates or variable seawater contributions to the fluid budget.Most of the micritic samples in this study show no Ce anomalies or, in few samples, real negative Ce anomalies (Fig. 6B).These data are indicative of precipitation under variable redox conditions as frequently observed in modern and ancient seep carbonates (Feng et al., 2009a;Jakubowicz et al., 2015;Wang et al., 2018;Wang et al., 2014).Negative Ce anomalies in seep carbonates may derive from episodic oxic conditions and seawater influx during carbonate precipitation (Feng et al., 2009a) or from organic matter complexation (preferential uptake of dissolved Ce 4+ by organic compounds) (Himmler et al., 2010;Kim et al., 2012).Although we cannot entirely rule out the complexation effect, the Mo/U EF ratio of sample L5 (2.8) is very close to the seawater ratio (3.1), and the first interpretation is more convincing.Most rim cements show positive Ce anomalies that can be interpreted to reflect precipitation under reducing conditions (Fig. 6B).Apparent Eu anomalies may result from analytical artefacts related to the interference of barium oxides forming during ICP-MS measurement (Dulski, 1994;Jarvis et al., 1989;Qi et al., 2005).
However, we rule out this possibility because there is no correlation between Eu/Eu* and Ba/Sm in micrites (Fig. 8B).On the other hand, rim cements are more susceptible to this artifact due the low Sm concentrations and high Ba content and show a strong positive correlation (R 2 = 0.9) (Fig. 8B).The positive Eu anomalies in micrites likely reflect mixing of reducing porewater

Mo and U enrichments and covariation
The Mo and U contents in Corella samples are highly variable (Tab.3) and frequently exceed the reported average values of the upper continental crust (~ 2.7 mg/kg for U, 3.7 for Mo; Taylor and McLennan, 1985) and modern marine sediments (1-5 mg/kg both for U and Mo; Morford et al., 2009).Our results are similar to those reported for modern methane-derived carbonates and seep-impacted sediments (Chen et al., 2016;Hu et al., 2017;Li et al., 2016;Liang et al., 2017).In micrite samples, Mo varies between 0.6 and 34.4 mg/kg, similarly to other published data from seep carbonates showing concentrations up to 12.4 mg/kg (Liang et al., 2017).Some of the carbonates display evident authigenic enrichments with Mo EF up to 233.5.As far as we know, this is the highest published value for methane-derived carbonates, and is similar to the value of Mo EF =273.0 encountered in sediment profiles from active seep areas in South China Sea (Chen et al., 2016).High Mo EF values (>100), found in three of our carbonates (3M, 10 and 30) reflect an episode of intense sulfate-driven AOM and the establishment of sulfidic porewater conditions.Scott and Lyons (2012) reported two distinct Mo ranges, one for environments with hydrogen sulfide confined to pore waters and one with sulfidic bottom waters.
In the first case, Mo is about 10 mg/kg and rarely exceeds 20 mg/kg.In case of persistent euxinic bottom waters, enrichments have a much wider range with concentrations consistently >60 mg/kg up to 100 mg/kg (Scott and Lyons, 2012).In our samples, Mo concentration rarely exceeds 20 mg/kg and the highest value obtained (34.4 mg/kg, sample 3M) is below the 60 mg/kg threshold (Tab.3).We suggest that episodic sulfidic conditions were mainly restricted to pore-waters and did not steadily influence the bottom water.
Interestingly, arsenic shows some detectable enrichments (As EF >3).Similar enrichments have been reported in seep carbonates collected above a mud volcano site in the Gulf of Mexico (Hu et al., 2014) and were interpreted as generated by an efficient particulate shuttle process (Algeo and Tribovillard, 2009;Crusius et al., 1996;Emerson and Huested, 1991;Magyar et al., 1993;Morford et al., 2005).In this process, methane-rich brines, episodically reaching the bottom waters, may induce Mn and Fe oxyhydroxides formation in the water column efficiently

A C C E P T E D M A N U S C R I P T
scavenging Mo and As from the surrounding seawater (Tribovillard et al., 2013).This particulate eventually sinks to the seafloor, becoming progressively buried and reduced within the anoxic zone.Released elements then become available for incorporation into iron sulfides and authigenic carbonates (Hu et al., 2014;Tribovillard et al., 2013).Further evidences for the involvement of the particulate shuttle control process in trace metal scavenging has been reported for Pliocene seep carbonates in Taiwan (southern Western Foothills, Taiwan; Wang et al., 2019).These carbonates are characterized by enrichments in Mo, As, and Sb that correlate with excess iron.Interestingly, the same samples show high (Mo/U) EF ratios but U EF <2.Based on field constraints, they suggest that these enrichments are related to iron shuttle processes with no influence of deep ascending fluids.In our study, As EF values show a clear positive correlation with Mo EF (R 2 = 0.67; Fig. 8) but lack any trend with Fe/Al (R 2 < 0.1; Fig. 8).We suggest that, rather than seawater, the source of As were the upward migrating fluids.The Corella seep carbonates thus represent an example of structurally controlled seepage system where trace metal enrichments are not influenced by the particulate shuttle process.
Our data show a range of (Mo/U) EF between 0.7 and 9.7; when plotted in the Mo EF vs U EF diagram, they mostly fall between the 0.3×(Mo/U) SW and 3×(Mo/U) SW trend lines (Fig. 6).The data distribution looks similar for seep carbonates from the northern Congo fan (Hu et al., 2014), and suggests very dynamic redox conditions.High flux of hydrocarbon-rich fluids toward the seafloor likely generated a compressed vertical redox zonation within the sediments.Under such conditions, the (Mo/U) EF basically depends on the AOM rate: at relatively slow rate of hydrogen sulfide production, the threshold concentration for conversion of molybdate to thiomolybdate (Helz et al., 1996;Zheng et al., 2000) was unlikely to be achieved and therefore U accumulation

Conclusions
In this study, we characterized a carbonate outcrop formed in a fault-related seepage system (the Miocene Corella outcrop, northern Apennines, Italy) by radiogenic strontium isotopes ( 87 Sr/ 86 Sr), rare earth elements and Mo-U systematics to constrain fluid sources and reconstruct the seepage dynamics and redox conditions.Most of the samples formed in equilibrium with Langhian seawater, in agreement with the age of the host sediments.These findings confirm that the SIS methodology can be successfully applied to seep carbonates when the diagenetic alteration is negligible.
We also found evidence of the contribution of a deeper fluid source characterized by a highly radiogenic Sr isotope signature.This signal was recorded in carbonate phases precipitating in cavities and fractures and micrite from the conduit-rich facies at the base of the carbonate body, in conditions less influenced by seawater.This radiogenic signal likely derived from fluids that interacted with detrital clays of the underlying terrigenous turbidite successions, conveyed toward the seafloor by faults linked to the Apenninic deformational front.
Molybdenum and uranium enrichments in carbonates indicate an overall intense AOM and highly variable suboxic porewater conditions, episodically sulfidic.Although REE are very effective in representing temporal variations of redox conditions, they do not always adequately record the presence of a deeper fluid source due to mixing effects, and cannot effectively constrain the complexity of fault-controlled seepage systems.
With this study, we demonstrated that the integration of 87 Sr/ 86 Sr ratios and Mo-U systematics in ancient seep carbonates represents a novel and powerful approach to the investigation of fossil systems potentially allowing for more complete paleoenvironmental
7 and 9.7.Eight samples show (Mo/U) EF below that of seawater (~3.1, Algeo and Tribovillard, 2009), two samples (19, L5) has (Mo/U) EF ≃ (Mo/U) SW and 11 samples have (Mo/U) EF >(Mo/U) SW .The rim cements have Mo concentration from < 0.1 to 3.0 mg/kg and U from < 0.01 up to 1.4 mg/kg; Al concentrations below detection limits do not allow normalization of Mo and U contents and the calculation of respective enrichment factors.

(
positive anomaly) with variable proportions of seawater (no Eu anomaly).An additional factor to be considered is the possible contribution of deep-sourced fluids that generally carry a Euenriched signal generated by Eu fractionation at elevated temperature(Kim et al., 2012).Based on the geological context and our Sr isotope data, it is reasonable to assume that deeper have also influenced the REE budget and contributed, in part, to the development of positive Eu anomalies in porewater.
prevailed over Mo, causing (Mo/U) EF <3.1.Enhanced AOM rates linked to higher CH 4 concentrations, on the contrary, progressively promoted sulfidic conditions within the SMTZ, resulting in (Mo/U) EF >3.1.No correlation between Mo EF or U EF and 87 Sr/ 86 Sr (Fig.7) suggests that there is not a straightforward relation between the inception of strongly anoxic and/or the input of radiogenic fluids.The concomitant occurrence of Mo EF >3.1, MREE enrichments and absence of positive Ce anomalies in some carbonates reflects the highly dynamic redox conditions during carbonate precipitation and supports the hypothesis of a compressed vertical geochemical zonation where the Fe-reduction zone partially overlaps the SO 4 -reduction zone.
Fig. 1.A) Simplified geological map of the northern Apennines showing the main structural units (different colors) and location of the Corella seep carbonate outcrop (star symbol) (modified from Argentino et al., 2019).B) Distribution of authigenic seep carbonate bodies and stratigraphic and structural relationships with surrounding units.Carbonate bodies were named from L1 to L6. C) Structural sketch of the northern Apennine accretionary wedge/foredeep during the Middle Miocene.Corella seep located on the thrust-related anticline is indicated by the black star.

Fig. 2 .
Fig. 2. A) Microphotograph showing dark gray matrix micrite (mm) enclosing fine siliciclastic grains, cloudy rim cement lining a cavity (rc) and limpid blocky calcite (bc) filling the center of the cavity, not considered in this study as related to late diagenesis (sample 15).B) Cloudy calcite cement filling a conduit (cc) within the micrite.Micrite and conduit filling are cut by late sparry calcite (sample condA).Scale bar = 1 mm

Fig. 5 .
Fig. 5. Strontium isotope data for the examined authigenic carbonates plotted on the seawater Sr curve of McArthur et al. (2012).Most micritic samples show 87 Sr/ 86 Sr ratios corresponding to Langhian seawater Sr composition (specifically MNN5a biozone, red square).Sr isotope composition of the rim cements has no age implication, but here they are used to highlight deviations from coeval seawater.Note that the most radiogenic samples, 3M, 3V and L2COR1VB are out of scale, not included in this figure.

Fig. 6 .
Fig. 6.Rare earth element geochemistry of the Corella seep carbonates.A) PAAS-normalized REE patterns of micritic samples and calcite cements.B) Ce/Ce* vs. Pr/Pr* diagram of the Corella samples (after Bau and Dulski,1996).Field I: no Ce anomaly; Field IIa: unreal negative Ce anomaly caused by a positive La anomaly; Field IIb: unreal positive Ce anomaly caused by a negative La anomaly; Field IIIa: real positive Ce anomaly; Field IIIb: real negative Ce anomaly.Feld IV: positive Ce anomaly masked by a positive La anomaly.
. A lucinid shell sample (L2G) age is in agreement with the micrite Sr isotope-age, which means it formed at shallow depth, close to the seafloor.These results indicate that .Recently,   Zwicker et al. (2018)have argued that during recrystallization of aragonite to calcite, REEs do not necessarily behave conservatively if early diagenesis is influenced by evolving porewater composition due to different biogeochemical processes, as in the case of cold seep environments.
(Taylor and McLennan, 1985)le /(X / Al) PAAS ], samples are normalized using the Post Archean Australian Shale compositions(Taylor and McLennan, 1985).The authigenic fraction of Fe is expressed as Fe/Al ratio.n.d.: no data.