Chloride behaves conservatively during most diagenetic processes, making it a useful tracer for identifying the mixing of fluids from different sources (Martin et al., 1996). In contrast, many others solutes are directly altered by diagenetic reactions, thereby changing solute-to-chloride ratios (Martin et al., 1996). Consequently, chlorinity anomalies and solute-versus-Cl profiles are commonly used to distinguish between fluid sources and diagenetic alterations (Martin et al., 1996).

The Na⁺ and Cl^- profiles at the Gelendzhik and Heraklion MVs are curved upwards (Figure 2), indicative of upward fluid advection (Bohrmann et al., 2003; Aloisi et al., 2004). Either way, their chemical imprint indicates upward migration of evaporite-derived fluids. Chlorinity in deeper fluids is approximately 3–4 times higher than that of seawater (Figure 2), consistent with values previously reported at the OMVF (e.g. Haese et al., 2006; Behrendt et al., 2023).

At the Gelendzhik MV, the Na⁺/Cl^- ratios range between 0.85 and 1.06 (Figure 5) and Cl^- and Na⁺ show a strong correlation (0.995), suggesting addition of both ions through halite dissolution (Aloisi et al., 2004) within the Messinian salts (Dählmann and de Lange, 2003). Simirarly, at the Heraklion MV, Na⁺ and Cl^- are strongly correlated (0.993, see Supplementary material_Table S2) indicating a similar source.

Sulfate concentrations at both sites exceed the seawater value and correlate significantly with Na⁺ and Cl^- (both r>0.98, p<0.01) supporting an evaporitic origin. Sulfate enrichment in the OMVF has been attributed to gypsum dissolution (Böttcher et al., 1998; Omoregie et al., 2009), with δ³⁴S and δ¹⁸O signatures further pointing to a Messinian evaporitic source (Böttcher et al., 1998).

Although Messinian salts dissolution is evident at both Gelendzhik and Heraklion MV, K⁺ and Mg²⁺ concentrations indicate different diagenetic processes. At the Gelendzhik MV, high Ca²⁺ and Mg²⁺ concentrations likely reflect carbonate or gypsum dissolution, whereas strong enrichment of K⁺ (up to eight times seawater value, Figure 2) suggests involment of a different mineral reaction. One possibility is that fluids originated from salt differing from those at other MVs, K and Mg enrichment has been attributed to mixing with fluids sourced from evaporites other than Mg-rich salts (e.g. bischofite) (Behrendt et al., 2023). Alternatively, subsurface silicate weathering could account for the release of bicarbonate along with Ca²⁺, Mg²⁺, Na⁺, and K⁺ into the pore fluids (Wallmann et al., 2008; Scholz et al., 2013). This interpretation is supported by the strong positive correlation between K⁺ and silicate (r = 0.937), and the higher silicate concentrations at this site. While silicate weathering should increase DIC, no pronounced DIC enrichment is observed (Figure 3). However, DIC production during silicate weathering may be counterbalanced by consumption during methanogenesis, potentially masking the weathering signal, as silicate weathering can buffer a substantial fraction of CO₂ produced by methanogenesis (Torres et al., 2020).

In contrast, fluids at the Heraklion MV display Na/Cl ratios between 0.85 and 1.26 (Figure 5), with values exceeding unity in deeper samples, indicating excess Na⁺. Such high Na/Cl ratios have also been reported in other mud volcanos fields (Li et al., 2014). Elevated Na/Cl ratios, together with downcore K⁺ depletion (K/Cl< 0.01 below 10 cmbsf) and significant negative correlations between K⁺ and Na⁺, Cl^-, SO₄²^-, and silicate (see Suppementary material_Table S2), point to smectite–illite transformation prior to fluid ascent (Martin et al., 1996; Hensen et al., 2007). This reaction involves uptake of K⁺ and release of Na⁺, leading to relative Na enrichment and K depletion compared to seawater.

Fluid compositions at both MVs indicate mobilization from a deep source layer beneath the Messinian salts, producing hypersaline ascending fluids. At Gelendzhik, Mg–K enrichments indicate a deep evaporite-derived fluid modified during ascent either through mixing with K-rich salts or through water–rock interaction with silicate minerals. At Heraklion illitization has led to K⁺ depletion and Na⁺ enrichment. This interpretation aligns with previous studies suggesting that the highly saline fluids point to the same endmember across the OMVF (Behrendt et al., 2023) which undergoes various alteration processes during ascension, influencing their final fluid signatures.

Heraklion is the only site where smectite-to-illite transformation is evident, even in shallow pore fluids. This reaction begins at approximately 60 °C and is nearly complete by around 160 °C (Kastner et al., 1991). Assuming a geothermal gradient of 30–35 °C/km for the Olimpi field (Camerlenghi et al., 1995), these temperatures correspond to fluid mobilization depths of ~2 to 5 km, supporting a deep-seated fluid source. It is also worth noting that, the mud flows indentified during the LEVECO cruise, were inferred to originate from sediments no deeper than approximately 2 kmbsf (Panagiotopoulos et al., 2020). This points to a decoupling between sediment and fluid mobilization depth, in agreement with previous studies in the OMVF (Deyhle and Kopf, 2001) and other mud volcano systems (Li et al., 2014; Doll et al., 2023).

At the Moscow and Milano MVs, Na/Cl and K/Cl ratios remain close to seawater values (Figure 5). However, both Cl^- and Na⁺ concentrations decrease with depth, indicating mild pore-water freshening. Cl^- concentrations decrease to 536 mM and 584 mM at Moscow and Milano, respectively, compared to near-surface values 619 mM and 627 mM. Similarly, Na⁺ concentrations decrease from 530 mM and 533 mM at the surface to 470 mM and 508 mM at depth.

Clay minerals dehydration, particurarly smectite-to-illite transformation is considered the most significant dehydration process in marine forearc environments (Saffer and Tobin, 2011) and has been identified as a key control on pore water isotopic composition (Dählmann and de Lange, 2003)., The observed chlorinity decrease may therefore reflect smectite-to-illite transformation occurring at depth, prior to fluid ascent near the surface. However, the absence of K⁺ depletion and Na⁺ enrichment relative to Cl^- prevents a conclusive attribution of the freshening to this process.

Our results highlight pronounced spatial variability in pore fluid concentrations across the studied MVs. Previous work has shown that substantial heterogeneity can also occur within individual MVs, such as the Carlos Ribeiro MV (Vanneste et al., 2011) and the Milano MV (Haese et al., 2006). At Milano, both saline-enriched and freshened fluids were reported within a distance of only ~500 m, suggesting that single mud volcanoes may be fed by multiple fluid sources (Haese et al., 2006).

At Moscow and Milano MVs, the SO₄:Cl ratios are consistently lower than those of Eastern Mediterranean seawater (~0.051) (Thompson et al., 1931), except at the Gelendzhik MV (Figure 5). OSR and AOM-SR both consume sulfate and produce DIC, but with distinct stoichiometric relationships (Masuzawa et al., 1992; Kastner et al., 2008).

Organic matter oxidation linked to sulfate (OSR) can be described by Equation 1 (Masuzawa et al., 1992):

If CH₄ is present in sufficient concentrations, AOM-SR may become the dominant process for SO₄^2- depletion in the pore water (Hensen et al., 2003), with higher activity under increased methane fluxes (Lee et al., 2021). AOM-SR can be described by Equation 2 (Boetius et al., 2000):

To distinguish between these pathways, we calculated the Δ[DIC + Ca²⁺ + Mg²⁺] and ΔSO₄, where Δ represents the concentration difference between the overlying water and the pore water at a given depth (Masuzawa et al., 1992; Kastner et al., 2008). Correction for Ca²⁺ and Mg²⁺ accounts for carbonate precipitation effects (Masuzawa et al., 1992; Kastner et al., 2008). A 1:1 slope in this plot indicates that AOM-SR is the dominant sulfate-consuming process. A slope between 1 and 2 suggests that both SR pathways are occurring simultaneously, while a slope 1:2 indicates OSR (Masuzawa et al., 1992; Kastner et al., 2008).

At the Gelendzhik MV, nearly all data points fall below the 1:1 slope, with an estimated slope of approximately 0.63 (Figure 7). Although lower than the theoretical AOM-SR ratio, this likely reflects DIC-depleted fluids and/or DIC loss due to methane degassing during core recovery (Haese et al., 2003), consistent with AOM-SR dominance. At the Heraklion MV, most data points fall between the 1:1 and 2:1 slope, implying that both OSR and AOM-SR co-occur (Figure 7). At the Milano MV, OSR appears to be the dominant process for sulfate consumption, as data points fall at the 2:1 slope (Figure 7). At the Moscow MV, OSR also appears to dominate. However, several data points lie well above the 2:1 slope, indicating excess DIC. This excess is likely attributed to diffusion of DIC from deeper zones (Kim et al., 2011), consistent with evidence of fluid flow at Moscow MV based on radio-tracing methods (Tsabaris et al., 2020).

Stable carbon isotopes of DIC (δ¹³C_DIC) provide additional insights on sulfate-consuming processes and carbon cycling. δ¹³C_DIC represents mixing between multiple carbon sources, including bottom water DIC, DIC carried by ascending fluids, DIC produced through methane oxidation, and DIC produced via organic matter oxidation (Lichtschlag et al., 2010). Consequently, δ¹³C_DIC values depend on the relative abundance of these sources (Aloisi et al., 2000). If OSR is the sole process involved, δ¹³C_DIC values are expected to range between that of bottom water (~0‰) and that of sedimentary organic carbon. In this study, δ¹³C_org values ranged between -23.4‰ to -28.6‰ (Figure 6), indicating a mixture of pelagic organic matter (~ -21‰ in Eastern Mediterranean pelagic sediments (Fontugne and Calvert, 1992) and a more ¹³C-depleted fraction.

At the Milano MV, δ¹³C_DIC values showed a clear downcore decrease, reaching -29.6‰ (Figure 3). δ¹³C_DIC values lower than those of typical organic matter (~ -21‰) indicate a significant contribution from anaerobic oxidation of methane (Chen et al., 2022). AOM-SR is typically associated with δ¹³C_DIC decrease toward the SMTZ (Kim et al., 2011), as methane-derived carbon is markedly in ¹³C-depleted (Borowski et al., 1996; Whiticar, 1999). At the Heraklion MV, δ¹³C_DIC values also decrease downcore to ~ -25.2 ‰ (Figure 3), suggesting OSR dominance with a secondary contribution from AOM-SR, consistent with moderate methane concentrations and the stoichiometric ratios. In contrast, pore fluids at the Moscow MV exhibit less negative δ¹³C_DIC values, with a minimum of 17‰ at 26–28 cmbsf (Figure 3). Such values are consistent with a mixed DIC pool of OSR-derived DIC and upwardly diffusing DIC from deeper zones, where microbial reduction of CO₂ to CH₄ preferentially removes the lighter ¹²C, leaving behind the residual DIC pool relatively enriched in ¹³C (Whiticar, 1999; Chen et al., 2010; Kim et al., 2011). This interprentation is supported by the cation-adjusted ΔDIC to ΔSO₄^2- ratios, which indicate a contribution from deeper zones.

At the Gelendzhik MV, the δ¹³C_DIC profile exhibits a complex pattern, ranging from -11.4‰ near the sediment - water interface to more depleted values of -22.4‰ at 13.5 cmbsf, followed by a shift toward less depleted values at depth (16.1‰; Figure 5). The pronounced δ¹³C_DIC depletion at 13.5 cmbsf likely corresponds to an active AOM zone, as such zones typically exhibit strong ¹³C_DIC depletion (Whiticar, 1999). The reversal toward less depleted values at depth indicates influence of DIC contribution from deeper sources, consistent with stoichiometric evidence for dominant AOM-SR and high methane concentrations. These observations point to a significant contribution of ¹³C-depleted DIC from both AOM activity and upward diffusion from deeper sedimentary layers.

The depth of maximal δ¹³C_DIC depletion likely corresponds to an active AOM zone (Haese et al., 2003; Chen et al., 2010). To test this interprentation, an end-member mixing model based on chlorinity and δ¹³C_DIC values was applied. Using chlorinity end members of 631 mM (overlying seawater), 1512 mM (proposed active AOM-SR zone), and 1989 mM (deepest core sample at Gelendzhik), the pore fluid at 13.5 cmbsf is estimated to consist of ~ 35% overlying seawater and ~ 65% deep fluid. Using this mixing ratio, the expected δ¹³C_DIC value would be -14.5‰, whereas the measured value is -22.4‰. This additional -7.9‰ depletion is therefore attributed to in situ AOM activity, providing evidence for an active AOM zone at this depth.

At the Gelendzhik and Heraklion MVs, OSR and AOM-SR occur under hypersaline conditions. Similar findings have been reported from other Mediterranean MVs, such as Chefren and Napoli MVs, where AOM and OSR were observed at salinities up to 153 and 268‰, respectively (Omoregie et al., 2009). In the Tyro and Bannock brines (salinity ~260‰), sulfate-reducing bacteria were likewise active, indicating that extreme salinity does not necessarily inhibit organic matter degradation via sulfate reduction (Henneke et al., 1997). At the Mercator MV, AOM-related microorganisms thrived at salinities approaching halite saturation (Maignien et al., 2013). Beyond the Mediterranean, molecular evidence of AOM has been reported in hypersaline Gulf of Mexico sediments (Lloyd et al., 2006) and in continental hypersaline environments, such as the Salitral Negro lake, where diverse halophilic microbial communities are capable of anaerobic metabolism using various electron acceptors, including sulfate (Solchaga et al., 2022).

To distinguish in situ produced DIC from deep-sourced DIC, we estimated the isotopic composition of diagenetically added DIC to the pore-water pool using a δ¹³C_DIC x DIC versus DIC plot following Miller and Tans (2003). The slope represents the isotopic signature of the carbon source undergoing mineralization (Martin et al., 2000; Hu et al., 2010; Coffin et al., 2013; Wu et al., 2016). At the Gelendzhik MV, δ¹³C_added value was estimated at -53.1‰ (Table 2), indicating a strong contribution from AOM-SR (Coffin et al., 2013). At Milano MV, the δ¹³C_added was also depleted (-34.1‰), suggesting AOM-SR influence. In contrast, Heraklion and Moscow MVs had less negative values (-28.45‰, and -22.87‰, respectively), consistent with a dominant OSR signal (Figure 8). These results highlight spatial variability in diagenetically carbon sources: AOM-SR dominates at Gelendzhik, OSR at Heraklion and Moscow, while Milano reflects a mixed contribution.

To assess carbon sources, we examined the sedimentary organic carbon pool. Using the δ¹³C_org x C_org versus C_org plot (Figure 8), we estimated the isotopic signature that organic carbon within the sediments, would exhibit upon remineralization. Unfortunately, data for Milano MV are available. The estimated slopes were -28.8‰, -30.4‰, and -28.0 ‰ at the Gelendzhik, Heraklion, and Moscow MVs, respectively (Table 2), more depleted than typical Eastern Mediterranean pelagic sediments, suggesting contributions from methane-derived biomass (Aloisi et al., 2000; Haese et al., 2003). These values do not demonstrate strong spatial variability among the studied MVs.

By comparing the δ¹³C signature of diagenetically added DIC and sedimentary organic carbon, we assess whether DIC originates from in situ organic matter mineralization or from deeper sources. At Heraklion, the DIC added (-28.4‰) is close to Corg_added (-30.4‰), suggesting that DIC is mainly produced via local OSR. At Moscow, the DIC added (-22.8‰) is enriched relative to sedimentary organic carbon pool (-28.0‰), suggesting a significant component of deep-sourced, ¹³C-enriched DICfrom upward diffusion from methanogenic zones. At Gelendzhik, the large offset between DIC added (-53.1‰) and sedimentary organic carbon (-28.8‰) indicates a dominant AOM signal and in situ methane cycling.

These δ¹³C isotopic offsets at Gelendzhik and Moscow MVs highlight distinct carbon sources for pore fluids and sedimentary organic matter, indicating fluid migration and degassing from deeper methanogenic strata. Porewater geochemistry at these sites is influenced both by ongoing shallow diagenetic processes and upward transport of deep-sourced fluids and gases.