Pyrites in a salt marsh-ria system: quantification, morphology, and mobilization

Marine Geology 455 (2023) 106954 Available online 24 November 2022 0025-3227/© 2022 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/). Pyrites in a salt marsh-ria system: Quantification, morphology, and mobilization X.L. Otero a , b , * , P. Guevara a , c , M. S´ anchez d , I. L´ opez c , H.M. Queiroz e , A. Ferreira e , T.O. Ferreira e , G.N. N´ obrega f , R. Carballo d a CRETUS, Cross-disciplinary Research in Environmental Technologies, Departamento de Edafoloxía e Química Agrícola, Facultade de Bioloxía, Universidade de Santiago de Compostela, Spain b REBUSC Rede de Estaci´ ons Biol´ oxicas da Universidade de Santiago de Compostela, Estaci´ on de Bioloxia Mari˜ na A Gra˜ na, Ferrol, Spain c Departamento de Ciencias de la Tierra y la Construcci´ on, Universidad de las Fuerzas Armadas ESPE, Av. General Rumi˜ nahui s/n, P.O. Box 171-5-231B, Sangolquí, Ecuador d ´ Area de Ingeniería Hidr´ aulica, EPSE, Universidade de Santiago de Compostela, Lugo, Spain e Luiz de Queiroz College of Agriculture, Soil Science Department University of S˜ ao Paulo (ESALQ-USP), Av. P´ adua Dias 11, CEP 13418-900, Piracicaba, S˜ ao Paulo, Brazil f Departamento de Geoquímica, Instituto de Química, Universidade Federal Fluminense, Brazil ARTICLE INFO Editor: Dr. Adina Paytan Keywords: Salt marsh Soil Sediment Ría Pyrite Framboids ABSTRACT Galician Rias are among the most productive ecosystems in the world. Consequently, the soils of their salt marshes and sediments of the intertidal flats show high organic matter contents, reactive Fe, and sulfate, which promote pyrite synthesis and accumulation, using sulfate for organic matter decomposition. This work studies the morphological variability and concentration of pyrites (individual crystals and framboids) in different geochemical environments found in the Ria de Ortigueira (salt marsh soils and bottom sediments in the inner, middle, and outer section), addressing their dynamics in the marsh-ria system in relation to the hydrodynamic characteristics defined by tides and river discharges. Framboidal pyrites were the dominant morphology in marsh soils and sediments in the middle and inner sections of the Ria, while isolated crystals dominated its outer section. The results showed that lower marsh soils (colonized by Spartina) are the most favorable environment for pyrite synthesis, showing high pyritic Fe concentrations and high degrees of pyritization, largely exceeding the values observed in sediments from Galician Rias and from most sedimentary environments worldwide. However, the amount of framboidal pyrites present in the lower marsh (SPE: 4–5 ×10 4 framboids) was clearly lower than in bottom sediments of the inner and middle part of the Ria de Ortigueira (~2–7 ×10 6 framboids), mainly due to the fact that pyrites were found to form large framboids in lower salt marsh soils. Thus, the amount of framboidal pyrites does not seem to be a good indicator of redox conditions in modern marine sediments. Pyrite crystals found in the sediments of the Ria showed poorly defined vertices and facets, indicating their degradation and suggesting that a significant amount of the pyrites found in the middle and inner sections derive from marsh collapse. Finally, the output of framboidal pyrites towards the outer Ria de Ortigueira reflects the low intensity of residual flows in this Ria. Therefore, the pyrites observed in the outer section consisted only of isolated crystals, presumably formed in situ under low sulfate-reducing activity conditions. 1. Introduction Salt marshes and Rias are among the most productive ecosystems on Earth (Margalef, 1952; Odum and Odum, 1981). Both ecosystems show high primary productivity, which results in high organic matter contents in their soils and sediments (Howarth and Teal, 1979; Otero and Macías, 2002, 2003, 2010), but also act as sinks receiving significant contributions of inorganic materials (e.g., clay and silt) and dissolved elements from rivers and wastewater discharges (Filgueiras and Prego, 2007). The high organic matter contents in its soils and sediments are * Corresponding author at: CRETUS, Cross-disciplinary Research in Environmental Technologies, Departamento de Edafoloxía e Química Agrícola, Facultade de Bioloxía, Universidade de Santiago de Compostela, Spain. E-mail address: [email protected] (X.L. Otero). Contents lists available at ScienceDirect Marine Geology journal homepage: www.elsevier.com/locate/margo https://doi.org/10.1016/j.margeo.2022.106954 Received 24 June 2022; Received in revised form 16 November 2022; Accepted 22 November 2022 Marine Geology 455 (2023) 106954 2 related to the redox conditions imposed by the tidal variations and seasonal patterns in sea currents (Kirwan and Mudd, 2012). Under such conditions, anaerobic pathways of organic matter decomposition may prevail and promote pyrite formation and accumulation (Oenema, 1990; Souza et al., 2009). Several studies have pointed out that organic matter in salt marsh soils and sediments is mainly oxidized by reducing Fe oxyhydroxides and sulfate (Wiebe et al., 1981; Howarth, 1984, Canfield et al., 1992). The coupling of these two anaerobic pathways increases the concentration of Fe 2+ and reduced S forms (i.e., HS − , polysulfides) in interstitial waters and, ultimately, promotes pyrite (FeS 2 ) formation (Canfield et al., 1992; Otero and Macías, 2010). Under anoxic conditions, pyrite is considered the most stable final product of sulfate reduction (Howarth, 1984), but is also responsible for relevant ecological services, such as trace metal retention (Otero and Macías, 2003). In fact, pyrites significantly control the biogeochemistry of several elements within marine and intertidal sedimentary environments (Howarth, 1984). In strongly reduced environments, pyrite is formed with mackinawite (FeS mk ) as a precursor, and H 2 S acts as an oxidant in a kinetically fast reaction that can take place within days (Reaction 1; Rickard, 1970; Lan and Butler, 2014). FeS(mk)+H2S→FeS2+H2Reaction (1) Conversely, in environments with alternating redox conditions (e.g., salt marsh soils in higher physiographic positions), the diffusion of oxygen into the substrate promotes the precipitation of poorly crystalline Fe oxyhydroxides (e.g., lepidocrocite) and the formation of polysulfides (S 5 −2 ) due to the partial oxidation of H 2 S (Reaction 2). Under these oscillating redox conditions, pyrite is formed by two different pathways: (i) through FeS as an intermediate product (Reactions 4, 5), with slower kinetics (i.e., within years; Luther III, 1991; Rickard, 2012; Peiffer and Behrends, 2015), and (ii) by the direct precipitation from interstitial water occurring within days (Reaction 5; Howarth and Teal, 1979; Giblin and Howarth, 1984). 8FeOOH +5H2S→8Fe2++S2− 5+2H2O+14OH−Reaction (2) Fe2++SH−→FeS +H+Reaction (3) 4FeS +S2− 5+2H+→4FeS2+H2SReaction (4) Fe2++S2− 5+HS−→FeS2+S2− 4+H+Reaction (5) Although pyrite is a major sink for Fe and other trace elements (Berner, 1967; Huerta-Díaz and Morse, 1992; Otero and Macías, 2003), redox conditions may considerably affect its role as a metal scavenger. For example, pyrite microcrystals are stable under anoxic conditions and may experience regrowth once buried (Rickard, 2012) whereas, under oxic and near neutral to alkaline conditions (e.g., seawater), pyrite can be oxidized within short periods (i.e., days or weeks; Aller, 1980; Morse, 1991). In this context, the influence of sea currents, catastrophic events such as storms (Scheffer et al., 2001), and coastal erosion events (Bertness et al., 2005; Ruggiero, 2012; Taherkhani et al., 2020) can transport, redistribute, and oxidize pyrite crystals releasing the associated elements into seawater (Morse, 1991; Otero et al., 2005). Some of these elements are essential limiting nutrients for marine primary productivity (e.g., Cu, Co, Fe), while others are potentially toxic (e.g., As, Hg, Cd, Pb) to several living organisms. Thus, combining chemical extractions for Fe partitioning, electron microscopy for pyrite morphological analysis, and numerical hydrodynamic modeling, this study aims to improve our understanding of the connections between salt marshes and the surrounding sedimentary environments and their role in nutrient and pollutant transfer from the continent to the ocean. 2. Material and methods 2.1. Description of the study area and sampling procedures The Ria de Ortigueira and the studied salt marshes are located in the NW Iberian Peninsula (latitude UTM: 4841094 N, longitude UTM: 593782 E, Zone 29 T, Fig. 1), showing a mesotidal regime, with a maximum range of 4.5 m. In recent years, the progressive collapse of salt marsh channels and edges has been observed, the resulting material falling directly into the inner and middle sections of the Ria or being transported from the channels into the Ria (Fig. 1). The freshwater input is small, with a mean discharge of approximately 8 m 3 s −1 during the winter and 3 m 3 s −1 in the summer. Two soil cores were drilled from the Esteiro salt marsh: the first located in a lower physiographic position and dominated by Spartina maritima (cordgrass), and the second located a drainage channel. Next, one soil core was drilled in the Mera salt marsh, located in a higher physiographic position and covered by Spartina maritima, and, finally, one soil core was drilled in the Ladrido salt marsh occupied by Juncus maritimus (sea rush), also in a higher physiographic position (Fig. 1). Cores were collected using 50-cm-long polyvinyl chloride tubes (PVC) with an internal diameter of 11 cm. After sampling, soil cores were transported to the laboratory in a vertical position and under refrigeration (~4 ◦C). In the laboratory, samples were removed from the tubes and sectioned into 3-cm-long sections for the upper 15 cm and 5-cm-long sections for the deeper portion to obtain subsamples for analysis. The following salt marsh subsamples were then selected for analysis: Esteiro salt marsh dominated by Spartina maritima, 0–3 cm layer (referenced as SPE 0–3 cm) and 15–35 cm layer (SPE2 15–35 cm); drainage channel of the Esteiro salt marsh, 0–15 cm layer (0–15 CE cm); Mera salt marsh dominated by Spartina, 0–3 cm (SPM1 0–3 cm) and 15–45 cm layers (SPM2 15–45 cm); and Ladrido salt marsh occupied by Juncus maritimus, 0–3 cm (JL1 0–3 cm) and 15–35 cm soil layers (JL2 15–35 cm). Additionally, surface samples (0–10 cm) from the bottom sediment of the Ria de Ortigueira were collected using a gravity corer (n =117; Fig. 1) along the central channel, which was divided into three sections according to hydrodynamic conditions: a) inner section (IS), spanning from the mouth of the River Mera to Ortigueira county, b) middle section (MS), from the Esteiro salt marsh to the Caldeira inlet, and c) outer section (OS), from the Ladrido salt marsh to the open ocean shore (Fig. 1; Guevara et al., 2021). Samples were stored in plastic ziplock bags and transported to the laboratory in the dark at a temperature of ~4 ◦C. To study pyrite size and morphology, seven samples from the IS, five from the MS, and seven from the OS were analyzed. Subsamples of each soil and sediment sample were frozen at −20 ◦C for Fe partitioning and pyrite analysis, while the remaining samples were air dried at 45 ◦C. 2.2. Analytical methods 2.2.1. Characterization of hydrodynamics of the Ria de Ortigueira The hydrodynamics of the Ria was computed through highresolution numerical modeling using the Delft3D-FLOW model. This model solves the 3D Navier-Stokes and transport equations reading: ∂ u ∂ x+ ∂ v ∂ y+ ∂ w ∂ z=S(1) Du Dt =fv −g ∂ ζ ∂ x−g ρ 0∫z′=ζ z′=z ∂ρ ∂ xdz′+ υ h( ∂ 2u ∂ x2+ ∂ 2u ∂ y2)+ υ v( ∂ 2u ∂ z2)Dv Dt = − fu −g ∂ ζ ∂ y−g ρ 0∫z′=ζ z′=z ∂ρ ∂ ydz′+ υ h( ∂ 2v ∂ x2+ ∂ 2v ∂ y2)+ υ v( ∂ 2v ∂ z2)(2) ∂ p ∂ z= − ρ g(3) X.L. Otero et al. Marine Geology 455 (2023) 106954 3 Dc Dt =Dh( ∂ 2c ∂ x2+ ∂ 2c ∂ y2)+Dv ∂ 2c ∂ z2−λdc+Rs(4) Eq. 1 represents the mass conservation for an incompressible fluid; Eq. 2 stands for momentum conservation in the x and y directions; Eq. 3 represents momentum conservation in the z-direction, which, under shallow water conditions, turns into the hydrostatic pressure equation; finally, Eq. 4 represents the transport equation, which, in the present application, is solved for salinity and temperature conditions. In the mass conservation equation, u, v, and w express the velocity components in the x, y, and z directions, respectively, and S is the source term. In Fig. 1. A) Study area and sampling sites (salt marshes: red and white circles; bottom sediments: red circles). The Ria de Ortigueira sectors (IS - inner section; MS - middle section, and OS - outer section) were defined based on their hydrodynamic characteristics. B) Distribution of salt marshes systems in the ria de Ortigueira. C) Photograph showing the progressive collapse of the salt marsh edges. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) X.L. Otero et al. Marine Geology 455 (2023) 106954 4 addition, in the momentum conservation equations, t is time, g is gravity acceleration, f is the Coriolis parameter, ζ is the elevation of the sea surface for z =0, υ h and υ v are the horizontal and vertical turbulent eddy viscosity, respectively, ρ h and ρ 0 are the density and the reference density of ocean water, respectively, and p is pressure. Finally, in the transport equation, c is a constituent representing salinity or temperature, D h and D v are the horizontal and vertical turbulent eddy diffusivity, respectively, λ d is the decay process of the first order, and C s is the source term. More information about model implementation can be found in S´ anchez et al. (2014). As stated, the main objective of implementing the numerical model was to model the hydrodynamic capacity of the Ria to transport pyrite and to alter its distribution within the Ria system. To this end, the transient and residual flows were computed during a typical winter period, which is considered to have higher sediment transport capacity (due to the greater river discharge compared with the summer season). Furthermore, the residual flows were computed considering the action of tidal constituent M 2 (i.e., the largest lunar constituent) and winter river discharges. In particular, the model was applied during a springneap tidal cycle, using the following as boundary conditions: (i) main tidal constituents, (ii) oceanic characteristics (salinity and temperature), and (iii) river discharge in the inner Ria, considering the average values during the period of interest (Q =8 m 3 s −1 ). 2.2.2. Analysis of salt marsh soils and sediments During sampling, redox potential (Eh) values were measured using an oxidation-reduction potential platinum electrode with final readings corrected by adding the potential value (+244 mV) of a calomel reference electrode. The pH values were measured using a Crisol micropH 2000 pHmeter with a glass electrode previously calibrated with standard solutions (pH 4.0 and 7.0). Total organic carbon (TOC) contents were determined using a Leco TruSpec CHN device by calculating the difference between total C and inorganic C; the latter was determined on samples after removing organic C by calcination at 450 ◦C for 4 h (Cambardella et al., 2001). Total sulfur (S total ) was determined using a Leco SC-144DR analyzer. Finally, particle size distribution was determined by the pipette method (Gee and Bauder, 1986) after the removal of soil organic matter using 6% NaClO at pH 8.0 (Mikutta et al., 2005). All analyses were carried out on samples from the 117 sampling sites for bottom sediments in the channel of the Ria (0–10 cm) and on four salt marsh soil cores. A sequential extraction procedure was performed to obtain pyritic (Fe py ) and reactive Fe (Fe reactive ). Fe reactive was extracted using 20 ml of a 0.25 M sodium citrate +0.11 M NaHCO 3 solution and 3 g sodium dithionite (Canfield et al., 1992), whereas Fe py was extracted using 10 ml of concentrated HNO 3 after removing Fe silicate-associated (by treating with 10 M HF and shaking for 16 h) and organic matterassociated Fe (by treating with concentrated H 2 SO 4 ; see Huerta-Díaz and Morse, 1992). Both fractions (Fe reactive and Fe py ) were analyzed by atomic absorption spectroscopy. The degree of Fe pyritization (DOP), which allows for the comparison of pyritization under different Fe contents, was calculated according to Berner (1970; Equation 5). DOP (%) = (Fepy ) (Fepy +Fereactive)x100 (5) 2.2.3. Pyrite morphology and size Pyrites were concentrated from sediment samples and soil subsamples (i.e., SPE1 0–3 cm, SPM1 0–3 cm, JL1 0–3 cm, and JL2 15–35 cm) by the bromoform decanting method ( ρ =2.89 g cm −3 , Wilkin et al., 1996 ; ). For this, 9 g of freeze-dried samples were weighed, carefully disaggregated with a mortar, and moved to a decantation funnel; after shaking, they were left to decant overnight. The dense fraction deposited at the bottom of the decantation tube was washed with acetone and placed in an aluminum tray. Pyrite size and morphology were determined by scanning electron microscopy (SEM) (FESEM Zeiss Ultra Plus with EDX-Image) at a resolution of 0.8 nm and an accelerating voltage of 30 kV. The pyrites were divided into five categories: 1) individual framboids, 2) framboid clusters, 3) isolated crystals, 4) framboidal pyrites within diatom frustules, and 5) framboidal pyrites within plant tissues. Framboid sizes were assessed based on size distribution histograms using geometric parameters and log-normal histograms (Limpert and Stahe, 2011; Rickard, 2019b). Statistical parameters x‾ =arithmetic means and σ =standard deviation were subjected to geometric transformation using the parameter ( ω ), which allows for arithmetic transformation into geometric parameters. It is expressed as ω =1 +( σ / x‾) 2 (Limpert and Stahe, 2011; Rickard, 2019b), which allows calculating the geometric mean (x‾* =x‾ /√ ω ) and the geometric standard deviation σ * =exp. (√ ln ( ω )) (for further details, see Rickard, 2019a). 2.2.4. Distribution of framboidal pyrites in the salt marsh-ria system This calculation considered the volume of each sphere, the mass corresponding to that volume, and pyrite contents obtained from Fe py content, assuming that all Fe py was found in the framboidal form. Eqs. 6–8 were used for this calculation, additionally assuming that framboids were solid spheres without interstitial spaces; their diameter was assumed to be the previously calculated geometric diameter (Limpert and Stahe, 2011; Rickard, 2019b): Sphere volume :V(cm3)=4 3 π (x* 2)3(6) Mass of a pyrite sphere (g):mP=dP*V(7) No.of framboids (framboids g−1):Nframboides =Fepy*MWpy mp (8) Where: x‾*: mean geometric diameter of pyrites in each section of the Ria or each type of salt marsh, d p : pyrite density =5.1 g cm 3 ; Fe py : mean content of pyritic Fe in each section, and MW py : molecular weight of pyrite =120 g mol −1 . 2.3. Statistical analysis Descriptive statistics were performed using the Minitab17 software. Frequency histograms for framboidal pyrite sizes were built using a lognormal function (Rickard, 2019b), whereas relationships among variables were established using Spearman correlation tests. Significant differences in the characteristics of pyrite among sites (types of salt marshes or sections of the Ria) or depths (surface/deep) were determined by a one-way ANOVA, followed by a post hoc Holm-Sidack test, or a Mann-Whitney test, using the SigmaPlot 12.0 software. Furthermore, the comparisons among Ria sectors and salt marsh soils were also performed using discriminant analysis (DA), which develops a function that yields optimal discrimination of the Ria sectors (IS, MS, and OS) and salt marsh soils (SPE, SPM, CE, JL). Thus, through DA, the relative contribution of the variables to the separation of the groups can be identified, i.e., the most important variables for the sectors and salt marsh soils (Reimann et al., 2008). 3. Results 3.1. Transient flow dynamics and residual current The circulation at the IS and MS and the inner section of OS are tidedominated, evidenced by similar flows throughout the water column (Fig. 2), with slightly lower velocities in the bottom layers, reversed during flood and ebb tides. Concerning the middle and outer sections of OS, the greater depth and width result in a lower tidal circulation and a more significant contribution of river discharges to the local X.L. Otero et al. Marine Geology 455 (2023) 106954 5 Fig. 2. A) Transient circulation at mid-flood tide in surface (A1 and A2) and bottom (A3 and A4) layers during average spring tides and average winter freshwater discharges. B) Transient circulation at mid-ebb tide in surface (B1 and B2) and bottom (B3 and B4) layers during average spring tides and average winter freshwater discharges. C) M 2 tide-induced residual circulation and freshwater discharge in winter in surface (C1 and C2) and bottom (C3 and C4) layers. X.L. Otero et al. Marine Geology 455 (2023) 106954 6 hydrodynamics. However, the existing river discharges cannot significantly alter this area’s transient flow, which could also be described as a tide-dominated circulation. The highest flow velocities are found in MS, mostly ranging from 1.0 to1.5 m s −1 , but eventually reaching 2.0 to 2.5 m s −1 near its boundaries with OS. In IS, velocity is significantly lower, approximately 0.2 m s −1 , with exceptions reaching 0.7 ms −1 . Finally, in OS, the flow pattern shows significant variations. In its innermost part, close to MS, velocities are higher (up to 1.5 m s −1 ) while downstream, at the middle and outer areas, where the depth and width of the Ria increases, velocities decrease to values close to 0.1 m s −1 . As a result, intense wind action could significantly influence the hydrodynamics in OS’s middle and outer portions. The residual circulation patterns varied widely throughout the Ria. Vertically homogeneous downstream flows were present in IS, MS, and the inner section of OS (Fig. 2), although with slightly lower velocities in the bottom layers and a 3D flow in the middle and outer portions of OS. The largest magnitudes of residual circulation were found in the same areas as for transient circulation. Transient circulation was concentrated in a limited area corresponding to the outer sections of MS and the inner section of OS, with downstream velocities around 0.20–0.25 m s −1 in the surface layer and 0.15 m s −1 in the bottom layer. In this context, to determine the potential for pyrite transport in the different areas, residual flows were computed considering the action of tidal constituent M 2 and winter river discharges (Fig. 2). The overall pattern corresponds to a positive estuarine circulation, evidencing the greater importance of river discharge over the tide for generating net circulation in this area. Finally, concerning IS, residual velocities are weak throughout the water column (~0.01 m s −1 ), one order of magnitude lower than in MS. 3.2. General characteristics of sediments and soils Bottom sediments from the three sections of the Ria were dominated by sand (average sand contents ranged from 71 to 93%; Table 1), increasing from the inner to the outer section. Texture in the IS and MS was highly variable, with samples showing fine fractions (i.e., silt and clay) accounting for 90%, whereas other sites showed sand contents higher than 95%. Sediment pH ranged from 6.9 to 7.9, and redox conditions ranged from strongly reduced (Eh <+100 mV) to oxic (Eh >+400 mV; Table 1). On average, IS showed significantly lower Eh (compared with MS and OS (Table 1). TOC and S total contents decreased from IS (TOC: 1.75 ±1.28%; S total :0.54 ±0.43%) to OS (TOC: 0.57 ±0.50%; S total :0.15 ±0.07%; Table 1). Salt marsh soils showed a finer and more homogeneous grain size composition than sediments, with a dominance of fine fractions (mean values ranging from 89% to 98%; Table 1). The pH values were circumneutral, with the lowest values found in the upper layer of SPE soils (i.e., the lower salt marsh; pH at SPE1 0–3 cm =6.2; Table 1). SPE soils also showed the highest Eh values (SPE1 0–3 cm: +459 mV), showing a decreasing trend at greater depths (SPE2 15–35 cm, Eh = +35 ±123 mV; Table 1), indicating an anoxic environment. The remaining soils occupied by cordgrass in the higher salt marsh (SPM) and the channel (CE) showed the dominance of anoxic conditions (average Eh <+100 mV; Table 1), with higher pH values (average: 6.8 ±0.1 and 7.1 ±0.2 for SPM and CE, respectively). TOC was >6% in all salt marsh surface samples (0–3 cm) influenced by the presence of plants, with a marked decrease with depth (TOC in the 15–35 cm depth <6%) and with the CE samples showing the lowest contents (Table 1). A different behavior was observed for S total contents. In soils colonized by cordgrass (SPE and SPM), S total contents increased with depth, from 0.48% in the upper 0–3 cm to 3.06 ±1.64% at 15–35 cm in SPE, and from 1.42% to 1.97 ±0.48% in SPM. Sea rush soils (JL) showed lower S total contents, with lower values at deeper layers (Table 1). 3.3. Geochemical forms of Fe and DOP The bottom sediments from the Ria showed Fe reactive concentrations 3 to 25-fold higher than those of Fe py in the three sections (Table 2). IS showed the highest Fe reactive (202 ±56.1 μ mol g −1 ), Fe py (71.9 ±50 μ mol g −1 ), and DOP (23.8 ±14.8%) contents, compared with the other sections (Table 2). In MS, Fe reactive and Fe py values were 137 ±70.4 μ mol g −1 and 11.3 ±20.1 μ mol g −1 , respectively, with a DOP value of 5.6 ±8.1% (Table 2). In OS, Fe reactive (74.8 ±28.5 μ mol g −1 ), Fe py (2.4 ±2.3 μ mol g −1 ), and DOP (2.6 ±1.6%) were clearly lower (Table 2). Considering the results observed by the discriminant analysis (DA, Fig. 3), higher Fe pyrite and DOP were highly related to higher S total and TOC contents, which were associated with IS samples but inversely related to Eh and sand. On the other hand, MS and OS were associated with higher Eh and sand contents and lower S total , DOP, TOC, and Fe pyrite . The differentiation between MS and OS zones could be better explained by higher pH values observed at MS (Fig. 3A). The highest Fe py contents were found in SPM2 15–45 cm, followed by CE, SPE2 15–35 cm, SPM1 0–3 cm, SPE1 0–3 cm, JL1 0–3 cm, and JL2 Table 1 Properties and composition of bottom sediments and salt marsh soils from the Ría de Ortigueira (mean ±SD). Sample pH Eh Sand Clay+Silt TOC S total mV - Bottom sediments IS (n =64) 7.3 ±0.2 +112 ± 143 71 ± 25 29 ±25 1.75 ± 1.28 0.54 ± 0.43 MS (n =28) 7.6 ±0.1 +314 ± 137 89 ± 25 11 ±28 0.92 ± 1.42 0.19 ± 0.19 OS (n =25) 7.4 ±0.1 +311 ± 134 93 ± 10 7 ±10.8 0.57 ± 0.50 0.15 ± 0.07 Low salt marsh soils SPE1 0–3 cm (n =1) 6.2 +459 1.8 98.2 6.31 0.48 SPE2 15–35 cm (n =4)* 6.0 ±0.2 +35 ± 123 7.1 ± 7.7 92.9 ± 7.7 5.30 ± 1.04 3.06 ± 1.64 Salt marsh channel CE (n =10) 7.1 ±0.2 +97 ± 114 10.6 ±1.8 89.4 ± 1.8 1.63 ± 1.02 0.79 ± 0.52 High salt marsh soils SPM1 0–3 cm (n =1) 6.8 -39 2.6 97.4 7.31 1.42 SPM2 15–45 cm (n =4)* 6.8 ±0.1 -85 ±6 3.3 ± 0.9 96.7 ± 0.9 4.62 ± 0.29 1.97 ± 0.48 JL1 0–3 cm (n =1) 6.3 +456 4.3 95.7 9.73 0.40 JL2 15–35 cm (n =4)* 6.3 ±0.1 +159 ± 17 7.4 ± 1.7 92.6 ± 1.7 3.58 ± 1.84 0.16 ± 0.08 Table 2 Reactive (Fe reactive ) and pyrite Fe (Fe py ) and degree of Fe pyritization (DOP) in bottom sediments and salt marsh soils from Ria de Ortigueira (mean ±SD). Sample Fe reactive Fe py DOP μ mol g −1 % Bottom sediments - Ría IS (n =7) 202 ±56.1 71.9 ±50 24 ±15 MS (n =5) 137 ±70.4 11.3 ±20.1 5.6 ±8.1 OS (n =7) 74.8 ±28.5 2.4 ±2.3 2.6 ±1.6 Low salt marsh soils SPE 0–3 cm (n =1) 239 21.1 8.1 SPE 15–35 cm (n =2) 77.4 ±37.0 231 ±27.6 75 ±11 Salt marsh channel CE (n =3) 73.0 ±29.0 263 ±104 76 ±16 High salt marshes soils SPM 0–3 cm (n =1) 200 81.0 29 SPM2 15–45 cm (n =2) 64.2 ±12.7 320 ±40.3 83 ±1.0 JL 0–3 cm (n =1) 253 2.1 0.8 JL 15–35 cm (n =2) 292 ±68 1.3 ±0.2 0.4 ±0.2 X.L. Otero et al. Marine Geology 455 (2023) 106954 7 15–35 cm (Table 2). In salt marsh soils, DOP ranged from very low values in JL2 (0.6 ±0.2%) to very high values in the lower salt marsh (SPE2 15–35 cm; 75.1 ±11.2%) and soils of Spartina-dominated higher salt marsh (SPM2 15–45 cm, 83.4 ±1.0%, Table 2). Deeper soil samples from SPM2 15–45 cm, SPE2 15–35 cm, and CE channel samples showed higher DOP levels (DOP >75%), compared with those found in SPM1 0–3 cm, SPE1 0–3 cm, and JL and JL2 (Table 2). The DA showed a clear differentiation between CE and the plant-covered sites (SPE, JL, and SPM), with higher pH, sand, Fe py, and DOP values. In contrast, plantcovered sites showed higher TOC, Fe reactive , and S total (Fig. 3B). Comparing plant-covered sites, SPM was associated with lower TOC and higher sand contents and pH, whereas JL was mostly associated with higher reactive Fe, and SPE was mostly associated with higher Eh and S total . 3.4. Pyrite distribution and morphology A total number of 1036 pyrites were observed by SEM in sediment and soil samples (Table 3). Individual framboidal pyrites (578 observations) and individual pyrite microcrystals (219 observations) were the dominant morphologies, followed by framboidal pyrites within organic structures (such as diatom frustules, 68 observations), whereas pyrites within plant tissues were less common (34 observations; Fig. 6, Table 3). Bottom sediments showed notably lower pyrites (319 observations) than salt marsh soils (717 observations). In the bottom sediments, framboidal pyrites and individual microcrystals were the most commonly observed category (Fig. 4, Table 3). Moreover, their Fig. 3. A) Discriminant analysis for the sections of the Ria (IS – Inner section, MS – middle section, OS – outer section) and B) salt marsh soils with sediment and soil characteristics. Table 3 Pyrite morphologies in the studied environments of the Ria de Ortigueira. Sample Single crystals Framboids Clusters Plant tissues Diatom frustules Bottom sediments IS 0 112 0 9 0 MS 158 15 0 4 0 OS 21 0 0 0 0 Low salt marsh soils SPE 0–3 cm 0 12 8 0 0 SPE 15–35 cm 0 65 117 11 0 salt marsh channel CE 38 52 2 10 1 High salt marsh soils SPM 0–3 cm 0 163 1 0 12 SPM2 15–45 cm 2 150 9 0 55 JL 0–3 cm 0 7 0 0 0 JL 15–35 cm 0 2 0 0 0 Total 219 578 137 34 68 X.L. Otero et al. Marine Geology 455 (2023) 106954 8 morphology showed poorly defined facets, edges, and vertices (Fig. 4). MS and OS showed poorly defined pyrite microcrystals with cubic, pyritohedral (dodecahedral), and octahedral habit (Fig. 4), whereas IS samples also showed framboidal pyrites incorporated into plant tissues (Fig. 4E and F: Table 2). Framboids were only observed in IS and MS, while at OS, a low number of isolated pyrite microcrystals were observed (21 observations, Fig. 4J and K). A greater number of pyrites were counted in salt marsh soils, corresponding to 78% of the total number of observations, as well as a wider diversity of sizes and morphologies (Table 3; Fig. 5). For example, Fig. 4. Microphotographs of pyrites from the bottom sediments of the Ria de Ortigueira. Framboidal pyrites from the inner section (A-F). Pyrites filling plant tissues (E and F). Pyrites from the middle section (G-I) and from the outer section (J-K), where the dense fraction is formed mainly by siliciclastic sediments (J) and very low content of pyrites (K). Pyrites from the bottom sediments showing octahedral (A, I) and cubic (B) habits. However, unlike the pyrites observed in marsh soils (Fig. 6), pyrite framboids in the sediments of the Ria are formed by microcrystals with poorly defined facets, edges, and vertices and cavities indicating that the framboids have been under unstable conditions. X.L. Otero et al. Marine Geology 455 (2023) 106954 9 in addition to isolated framboids and crystals, polyframboids/clusters formed by the aggregation of several framboids were commonly found in salt marsh soils, as framboids formed in diatom frustules and plant tissues (Fig. 5). Additionally, larger framboids and microcrystals with well-defined facets and edges were observed in salt marshes, especially in the deeper soil layers colonized by Spartina maritima (both in the higher and lower salt marsh, i.e., SPE2 15–30 and SPM2 15–45 cm). Most of the framboids observed in salt marsh soils were formed by tens to hundreds of pyrite microcrystals with well-defined habits (Fig. 5). However, in some cases, particularly in the upper layers of soils under Spartina maritima and Juncus maritimus, microcrystals showed clear signs of alteration, such as poorly defined edges and facets or clear evidence of dissolution or corrosion (Fig. 6). 3.5. Size distribution of framboidal pyrites Individual framboid sizes ranged between 2 and 85 μ m (Table 4). The highest geometric mean (x‾*) corresponded to sample SPM2 15–35 cm (x‾* =54.8 μ m), followed by samples from JL2 15–35 cm (x‾* = 37.7 μ m), SPE2 15–35 cm (x‾* =35.8 μ m), and SPE1 0–3 cm (x‾* = 21.9 μ m). Sediments from the Ria and the salt marsh channel (CE) showed smaller framboids (<10 μ m). Most framboids (44%) were medium-sized (3–30 μ m) and mainly associated with organic structures, such as diatom frustules and plant tissues (Table 4; Fig. 7). Large framboids (i.e., those with diameters >30 μ m; Rickard, 2019a, 2019b, Wilkin et al., 1996) accounted for 34% and were only found in salt marsh soils (Fig. 7). 4. Discussion 4.1. Size, morphology, and stability of framboids in salt marsh soils and sediments Framboids are the main pyrite morphology in modern sedimentary environments, such as marine sediments and coastal marsh soils (Wilkin et al., 1996; Wilkin and Barnes, 1997). The size of framboidal pyrites observed in sediments from the Ria follows the range of pyrites observed in most marine sediments, i.e., from 5 to 10 μ m (Wilkin et al., 1996; Roychoudhury et al., 2003; Rickard, 2012), whereas pyrites found in salt marsh soils were much larger (arithmetic mean values ranging from 10.2 Fig. 5. Microphotographs of pyrites from salt marsh soils. Pyrites from lower salt marsh soils (SPE) covered by Spartina maritima (A-C). Framboids from SPE 0–3 cm formed by spherical framboids of variable sizes (A). Large pyrite polyframboids (>100 μ m) from the upper portion of soil in SPE1 0–3 cm (B). Individual pyrites microcrystal with well-defined surfaces (C). Pyrites from higher salt marsh soils (SPM) colonized by Spartina maritima (D – F). Framboid formed by pyrite microcrystals with a well-defined dodecahedral habit (D). High abundance of framboids and filled diatom frustules in SPM soil (E). Large framboid cluster within plant tissues in SPE2 15–35 cm soil (F). Pyrites from higher salt marsh soils colonized by Juncus maritimus (Sea rush, JL samples) (G-I). Individual framboidal pyrite from salt marsh soil JL1 0–3 cm (G). Framboidal (H) and polyframboidal (I) pyrites from the deep layer of JL2 (15-35 cm) with poorly defined crystals. X.L. Otero et al.