1. Introduction
Metal(oid)s are pervasive in all earthly environments, some (e.g., Fe, Ni, Cu and Zn) are essential for plant growth and human and animal health. However, some become highly toxic to the surrounding ecosystems when present in excessive amounts [
1,
2,
3]. Metal(oid)s can enter into various surface and subsurface environments from both natural (via dissolution of metal(oid)-bearing minerals) and anthropogenic sources (waste materials, dredged materials, biosolids, fly ash, and atmospheric deposits) [
1,
4,
5,
6]. Consequently, they are marked as serious ecological pollutants due to their persistence and non-degradability in the environment [
1,
2,
7,
8,
9,
10,
11,
12].
By connecting land and sea, salt marshes (hypersaline environments) are among the most widespread, productive and vulnerable coastal ecosystems, providing a wide range of benefits to coastal populations and most ecological services such as: shoreline protection, fisheries support, water quality enhancement, maintenance of healthy marine ecosystems, habitat provision, wildlife conservation, key sink/source of organic material, nutrients and carbon sequestration [
13,
14,
15]. Unfortunately, salt marshes also serve as repositories for pollutants from terrestrial runoff, e.g., nutrients, pesticides, halogenated hydrocarbons and metal(oid)s [
14,
16,
17,
18,
19,
20,
21]. Salt marsh sediments are therefore highly enriched in metal(oid)s due to salinity, tides and occasional flooding [
22,
23,
24,
25]. They represent not only a sink, but also a source of metal(oid)s for all ecosystems (water, plants, animals and also humans) surrounding the salt marshes [
16,
17,
20,
23,
24,
25].
Depending on the chemical and geological conditions in the sediments, different forms of metal(oid)s are associated with a variety of inorganic and organic solid phases and dissolved in the pore water of the sediment [
1,
4,
26,
27]. In general, metal(oid)s are mainly concentrated in the fine-grained size of the solid phases [
1,
4,
28,
29]. Iron and manganese oxides/oxyhydroxides [
30,
31,
32,
33,
34] and metal sulphides [
31,
34,
35] are well-documented scavengers of metal(oid)s. Clay minerals are also one of the major carriers of metal(oid)s as they have enhanced sorption capacity especially during mobilization and diffusion processes [
36]. Metal(oid)s can be bound to various forms of organic matter: living organisms, detritus and coatings on mineral particles, etc. [
27,
34,
37].
In addition, early diagenetic release of metal(oid)s from sediments has been documented in studies of shallow coastal systems [
38,
39,
40,
41,
42,
43]. The diagenesis process, can cause solutes to become mineral saturated, resulting in the precipitation of authigenic sulphide, carbonate and phosphate minerals. These precipitates can sorb or co-precipitate metal(oid)s, and thus act as long-term sinks for pollutants in sediments. For example, Parkman et al. [
43] and Pirrie et al. [
44] demonstrated that early diagenetic sulphides act as sinks for Cu, Pb and Zn.
There are few studies addressing the mineralogy and chemistry of sedimental (detrital) grains and diagenetic mineral precipitates in coastal/salt marsh and—hypersaline environments [
45,
46]. The integrated mineral and chemical approach presented is necessary to identify the host species of metal(oid) minerals and the conditions that control the release of metal(oid)s into the system.
Understanding metal(oid) mineralogy in sediments is critical to predict and define metal(oid) translocation paths in target ecosystems. In terms of bioavailability (a prerequisite for translocation processes), a thorough understanding of metal(oid) associations in sediment solid components is necessary—whether they are bound to mineral constituents (and which mineral constituents) or organic matter. This is the first step to a comprehensive understanding of bioavailability and the transfer of metal(oid) to the system. Therefore, the following research objectives are pursued:
Determine detailed mineralogy (at the submicron level) and the chemical composition of hypersaline sediment particles.
Characterisation of metal(oid)-bearing minerals and their origin (geogenic and/or anthropogenic sources).
To assess the metal(oid) distribution in the hypersaline system.
2. Materials and Methods
The Sečovlje Salina Nature Park (
Figure 1) is located in the southwestern part of the Slovenian Adriatic coast, on the border with Croatia. It is the largest (an area of about 750 hectares) of the Slovenian coastal marsh wetlands and represents an ecologically unique environment with diverse ecosystems, including transitional forms between marine, brackish, freshwater and terrestrial ecosystems. These extreme environments of high salinity and aridity provide shelter to rare animal and plant species. Thus, salt-resistant or salt-tolerant plants (halophytes), also thrive in this saline environment. The area was designated a nature park in 1990. In 1993, it became the first Slovenian wetland to be included in the list of internationally important marshes under the auspices of the Ramsar Convention [
47,
48]. The Sečovlje salt-pans are among the few still active salt-pans in the Mediterranean and the first records of their operation date back to the 12th century. Since 2001, the Sečovlje salt-pans have been protected as a Nature Park by a special decree of the government of the Republic of Slovenia [
47,
48].
The Sečovlje Salina Park is located at the mouth of the Dragonja River into the Bay of Piran (Gulf of Trieste, northern Adriatic Sea). The catchment area of the Dragonja River is about 96 km
2 large and composed predominantly of Eocene flysch deposits characterised by a thin to medium bedded siliciclastic and carbonate-siliciclastic turbidite sandstones, marls, and meter-thick beds (megabeds) of calciturbidites. Cretaceous and Paleogene limestones are only found locally, near the outflow of the Dragonja River into the sea. Sediments in the Sečovlje Salina area originate from the geological hinterland of the Dragonja River [
23,
49,
50,
51].
The catchment area is basically a much closed geological area mainly towards the inland and thus we cannot find any anthropogenic sources from inland direction. The only possible pollution path from the inland could be through the air particles deposition—like this we can justify the presence of cinnabar or mercury particles originated from Idrija historical mine, which is almost 70 km by air from the Sečovlje Salina. Cinnabar or mercury small particles (e.g., μm-nm size) emitted into the air can travel thousands of miles in the atmosphere before they settle into the ground.
Conversely, we could trace the pollution from the sea side much more clearly. There are two large ports near the Sečovlje Salina: (1) Koper at the air distance of 14 km with cargo traffic and (2) Trieste at the air distance of 23 km with cargo and passenger traffic. On the other side of the Gulf of Trieste bay at the air distance of 28 km, is the mouth of the Soča/Isonzo River, which transports/transported different geogenic and anthropogenic material from inland, including the material from now long closed Idrija Hg mine and several small Cu deposits. However, the inflow of presented material into Sečovlje Salina area is strongly dependent on the currents in the Gulf of Trieste bay. In general, the currents are clock-wise (up along the Adriatic coast and down along the Italian coast), but it could be more complex, meaning that the currents could also bring the dispersed geogenic and anthropogenic material from Soča/Isonzo River, Trieste and Koper port.
Within the research area there are also many vineyards and orchards where CuSO4 was used as a pesticide spray.
In April 2020, samples of underlying sediment, its corresponding rhizo-sediment and upper
Salvia fruticosa plant were taken in the wider area of Sečovlje Salina (
Figure 2,
Table 1); (1–3) from the Piccia sampling site, where seawater evaporates, (4 and 5) from the crystallisation basin and (6) from the Poslužnica sampling site, which is a reservoir or collector of concentrated brine in the crystallisation area. At the time of sampling, the sampling sites were not covered with water (marine or brine water). Sediment and rhizo-sediment samples were collected for this study in order to comprehensively identify the mineralogical and geochemical signature and metal(oid)-bearing minerals in the samples studied. A representative sample of underlying sediment and a representative sample of rhizo-sediment for each sampling site marked in
Figure 2 consisted of 3 subsamples of underlying sediments and 3 subsamples of rhizo-sediment.
The underlying sediments were collected using plastic corers (a 15 cm long tube with an internal diameter of 10 cm) and packed in pre-cleaned plastic bags. The sediments were macroscopically homogeneous and bioturbated with a lighter grey colour due to oxidation. Samples of S. fruticosa plant and rhizo-sediment attached to the plant roots were also packed in pre-cleaned plastic bags. All samples were transported to the laboratory, where they were prepared for further analyses according to analytical protocols.
Underlying sediment (further expression used in the text: sediments (S)) samples were dried at 50 °C for 48 h. Ten grams of each sample was set aside for particle size analysis, and the remainder was sieved through a 0.315 mm polyethylene sieve to remove large organic and mollusk debris. For subsequent mineralogical and geochemical analysis, the samples (<0.315 mm) were quartered, milled and homogenised to a fine powder (<63 μm) using an agate mortar. For scanning electron microscopy energy dispersive spectrometer (SEM-EDS) analysis a portion of each quartered sample (<0.315 mm) was used to produce polished sections. The rhizo-sediment adhering to the plant roots was carefully washed with distilled water in a plastic container and left until all the particles had settled. The water was then removed and the samples were dried at 50 °C for 48 h. Further preparation of the rhizo-sediment samples was identical to that of the sediment samples. Two representative samples (15 g of sediment and rhizo-sediment) for the preparation of the clay fraction (<2 μm) were dispersed together with 200 mL of distilled water using a kitchen blender and an ultrasound. Sodium pyrophosphate was added to the suspension as a dispersant to promote dispersion and prevent flocculation. The suspension was gently stirred with a plastic spoon and allowed to stand for 45 min. After this time (Stokes’ law), the supernatant with a particle size of 2 μm was pipetted from the area 1 cm below the surface. The entire procedure was repeated three times for each sample to decant almost the entire fraction below 2 μm. The supernatants from 3 cycles were placed in a separate container and dried in an oven at 50 °C for 24 h. This dry clay fraction was used to produce a circular (diameter 2.5 cm) SEM-EDS polished section. The total thickness of the polished section was 1 cm.
The Slovenian salt production company (Soline Pridelava soli, d.o.o.), which manages salt production in the Sečovlje Salina Nature Park, issued a sampling permit. Sampling had no negative impact on endangered or protected species living in the Sečovlje Salina Nature Park.
The granulometric composition of the sediment samples studied was analysed using a laser granulometer Particle sizer (laser granulometry) and dynamic image analyser Fritch Analysette 22–28 with a measuring range from 40 nm to 2 mm (laser granolumetry) and 20 μm to 20 mm (dynamic image analyser). The sample was ultrasonically dispersed for 2 min before analysis. Grain-size frequency distribution plots were generated for each sample to check whether the distributions were polymodal. Numerical parameters for mean, median, primary mode, standard deviation, skewness, and kurtosis, percentage of clay, silt and sand were derived from the distribution data using the software provided. The position (i.e., size in μm) of a secondary mode was determined from the grain-size frequency plots. The image of individual grains were recorded by using an Analysette 28 Imagesizer (Fritsch) with a size range of 2800 to 20 μm.
Sediment pH was measured in situ using a portable pH meter (EUTECH Instruments) and further premeasured in the laboratory according to ISO standard 10390:2005 (suspension of sediment in water). Total organic carbon (TOC) content was determined using an Elementar Vario Micro CHNS elemental analyser. Prior determination, freeze-dried powdered samples were acidified with 6 M HCl to remove inorganic carbon [
52]. The precision of the method was 3%. To calibrate the analytical system the Calibration Standard sulfanylamide (with theoretical values N—16.27%, C—41.82%, H—4.68%, O—18.58%, S—18.62%) was used.
The mineral assemblage of the samples was determined using X-ray powder diffraction (XRD) Philips PW3710, equipped with Cu Kα radiation and a secondary graphite monochromator. Data were acquired at 40 kV, and a current of 30 mA over a range from 2 to 70° 2θ, at a rate of speed of 3.4θ/min. The oriented clay mineral aggregates were prepared by a combination of ultrasound dispersion, salt removal by centrifugation (3 × 3 min at 2500 rpm) and the glass slide method. Afterwards, the samples were treated with ethylene glycol-solvated condition and exposed to the reagent vapour for at least 24 h at 70 °C. Diffraction patterns were validated with the X’Pert HighScore Plus diffraction software version 4.6 using PAN-ICSD powder diffraction files and the Rietveld method, a full-pattern fit method used to compare the measured and calculated profiles.
SEM analysis was performed on ThermoFisher Scientific Quattro S using a Schottky effect field emission electron source (FEG-SEM) equipped with an Oxford Instruments UltimMax 65 energy-dispersive spectrometer (EDS). Microscopy and chemical analyses were conducted at 10 mm working distance, 20 kV acceleration voltage and 10 nA beam current for large grains (>10 μm), while acceleration voltage was reduced (10–15 kV) for clay minerals and particles smaller than 10 μm to achieve better spatial resolution. The instrument was calibrated with pure silicon and cobalt standards for light and heavy elements, respectively. The error in wt.% concentration at the 2σ level ranged from 0.05–0.25 for light elements and 0.30–0.40 for heavy elements (
Table 2). The EDS spectra were acquired over a period of 60 s. Before starting the analyses, the samples were coated with a thin film of amorphous carbon to ensure the electrical conductivity of the sample and prevent charge build-up.
Major oxides and trace metals content was obtained by ICP emission spectrometry, after lithium metaborate/tetraborate fusion and dilute nitric acid digestion (BV Mineral Lab—Acme, Vancouver, BC, Canada). The quality of the laboratory tests and objectivity was assured by the use of neutral laboratory tests. The accuracy of the analytical method, estimated by calculating the relative systematic error between the measured and recommended values of the reference materials STD DS11 and STD OREA262, was in the range of 2–13% with a median value of 7%. The precision of the analytical method, expressed as the relative percentage difference (% RPD) between the sample S6 duplicate measurements was in the range of 0–7% with a median value of 3%.
Basic statistical parameters for each element and multivariate statistical parameters were implemented using Statistica VII and Grapher 8 statistical software. The dissimilarity between objects was calculated using the Euclidean distance, and the objects were clustered using both the average linkage and Ward’s method (Cluster analysis (CA)). The Euclidean distance was adopted as it provides a greater emphasis to larger differences between variables. Principal component analysis (PCA) was used to disclose mineralogical and geochemical correlations within the data set.
4. Discussion
The distribution and the fate of metal(oid)s in salt marsh/saline sediments cannot be understood without knowledge regarding the deposited mineral particles. The geological background with the associated flysch rocks represents the primary source for the sediments and the detrital particles trapped within the sedimentary material represents the particles of the background rocks. Further on, primary and secondary clay minerals tend to bind free ions from the solution. The properties of the deposited mineral particles have a tremendous influence on the binding ability of metal(oid)s to mineral surfaces or incorporation into the crystalline structure of newly formed authigenic minerals [
87,
88,
89]. Moreover, particle size distribution is one of the most important factors affecting the ability of sediment particles to accumulate metal(oid)s. Metal(oid)s contents in surface sediments generally increase with decreasing grain size because of the affinity of metal(oid)s to bind to finer particles such as clay minerals [
46,
90]. Organic matter in the sediments could also contribute to the formation of metal-organic complexes [
4,
34].
In order to confirm the geogenic and anthropogenic origin of the Sečovlje Salina sediment material, the values of metal(oid)s in the investigated sediment samples were normalised to the values of Al. The enrichment factors (EF) were calculated using the normalised content in sediments and the average normalized contents of representative rock samples from the Sečovlje Salina geological background.
The EF values (
Supplementary Materials) obtained were similar for most of the metal(oid)s determined in the samples and pointed to two main sources of the sediment particles in the investigated area. Since the median EF values calculated for Sn (0.23), Cd (0.46), Ni (0.64), Cr (1.2), Cu (1.2) and Pb (1.3) ranged from 0.2 to 1.3, this supports the assumption that weathering of rocks from the hinterland is the predominant geogenic source [
91] of material deposited in the Sečovlje Salina area. Hg EF values (0.95) pointed to the Idrija Hg mine and air deposition of cinnabar or mercury particles. As (2.57), Bi (2) and Zn (2.74) EF values were higher than 1.5, which according to Szefer et al. [
92] and Chen et al. [
93] reflect anthropogenic contamination of the metal(oid) in question, e.g., minor enrichment. However, if we consider the calculated EF values with respect to different sampling locations, the results are slightly different. Samples from the Piccia area reveal lower EF values (1.76) and the samples from the crystallisation basin and Poslužnica have higher values (2.94). The mentioned difference results from the fact that the samples from the crystallisation basin and Poslužnica contain more salt (halite minerals), in which traces of As have already been identified [
94]. The occurrence of Bi was already defined as anthropogenic due to its elevated industrial application and the proximity of the studied area to ports, while Zn is a very well-known essential element for humans and animals [
3], and is therefore also quantitatively abundant in any ecosystem.
PCA (
Figure 12) explained 84.63% of the data variance in the first two ordination axes, highlighting significant positive correlations between the samples from the crystallisation basin and Poslužnica due to their higher metal(oid) and Sr content. A pattern of individual correlations among metal(oid) and Sr distribution between the samples from crystallisation basin and Poslužnica is also recognised. Samples from Piccia (S1, RS1, S2, RS2, S3 and RS3) clearly show lower metal(oid) and Sr content, thus we can track down negative connections or no connections with the samples from crystallisation basin (S4, RS4, S5 and RS5) and Poslužnica area (S6 and RS6).
As, Bi, Cr, Cu, Pb, Sn, Sr, Pb and Zn form a large integrated group, nevertheless, we can emphasise and define the differences in origin, distribution and association of the elements within this group. As, Pb and Zn could be associated with Fe/Mn oxides and oxyhydroxides (goethite and hematite in our study) [
30,
89,
95,
96], or incorporated into crystal lattices of clay minerals (especially illite) [
4] and into halite minerals [
94]. Cu is generally closely bound to organic matter [
4] and could be adsorbed to aragonite [
97]. In the following order, Zn > Pb > Cu display very high sorption capacities for aragonite and calcite [
4,
97,
98], and hence the proximity and interrelation with Sr. Bi exhibits many chemical properties similar to those of As, and is generally accompanied by Pb ores in small amounts [
85]. These could be the reasons for its positioning near As and Pb, although we have identified its anthropogenic origin. The distribution of Cr is influenced by geogenic sources, but, in sediments it could be (especially Cr(III)) rapidly and specifically adsorbed by Fe and Mn oxides and clay minerals [
4]. This adsorption increases with increasing pH and organic matter content, which is the case in the Sečovlje sediments [
4]. Clay minerals are assumed to be the sources of Sn and the measured Sn contents is very low. We cannot ignore the presence of tributyltin oxide in the surrounding waters, which is widely used as an antifouling paint for wood preservation in boats and ships due to its biocide effect [
44,
88].
Cd, Hg and Ni are located as individual outliers (
Figure 12). Cd predominantly forms precipitates with carbonates, including biogenic particles, but Cd adsorption is strongly controlled by the presence of competing cations, such as divalent Zn and Cu [
4]. Cinnabar grains (HgS) are strictly anthropogenic in origin and originate from the Idrija Hg mine. Ni is incorporated into detrital sediment particles and its accumulation is closely governed by Mn oxides and oxyhydroxides [
4,
89].
Whereas the investigated sediment samples are dominated by slightly oxic conditions (an active oxic/anoxic sedimentary boundary) and neutral to low alkaline pH, the precipitation of Fe and Mn oxides/oxyhydroxides, the formation of metal-organic complexes and framboidal pyrite production are preferred. Consequently, this also affects the metal retention. Changes in pH and salinity would undoubtedly alter the influence of the relative role of the absorption phases. For instance, an increased pH would favour the ionisation of carboxyl functional groups on organic matter and surface hydroxyl group on oxides [
45,
96] and releasing of metal(oid)s into the surrounding system, on the other hand, pyrite production would not occur.
5. Conclusions
This study demonstrates that understanding the origin and incorporation of metal(oid)s within different solid sediment particles (such as detrital particles, anthropogenic particles, clay minerals, organic matter and diagenetic minerals) is critical to identify their distribution, accumulation, retention and further transfer into a vulnerable and ecologically important ecosystem such as salt marshes.
The results suggest that the detrital particles from the geological hinterland, e.g., the type of rock fragments and minerals, the various anthropogenic sources, the particle size relative to the amount of clay minerals, the abundance of organic matter and early diagenetic precipitates, significantly affect the metal(oid) accumulation and current retention in corresponding sediment and rhizo-sediment content of the Sečovlje Salina. The metal(oid)s, e.g., As, Bi, Cd, Cr, Cu, Hg, Ni, Pb, Sn, Sr, Pb and Zn display many differences in the origin, distribution and association within the sediment components.
According to the results of XRD, ICP-ES, SEM-EDS and various statistical analyses, the studied elements were mainly associated with Fe/Mn oxides and oxyhydroxides (As, Cr, Ni, Pb, Zn), incorporated into or adsorbed onto the crystal lattices of clay minerals (As, Cr, Pb, Sn, Zn), halite (As) and aragonite/calcite (Cd, Cu, Pb, Sr, Zn) and associated with organic matter (Cu, Pb and Zn). Only As, Bi, Hg and Zn were recognised as anthropogenic, although it would be difficult to determine anthropogenic sources for As and Zn, as their abundance is more likely due to processes occurring in the sedimentary basin. Traces of As were found in halite minerals, while Zn is an essential element for all living beings and is therefore also present in larger quantities. BiO industrial compounds are very common cargo found in the nearby ports and Hg presence is the result of almost 500 years of historical mining activity in Idrija.