**Sediment Dynamics of the Neretva Channel (Croatia Coast) Inferred by Chemical and Physical Proxies**

**Federico Giglio 1, Stefania Romano 2,\*, Sonia Albertazzi 2, Francesca Chiarini 2, Mariangela Ravaioli 2, Marco Ligi <sup>2</sup> and Lucilla Capotondi <sup>2</sup>**


Received: 12 December 2019; Accepted: 21 January 2020; Published: 23 January 2020

**Abstract:** We examined the transport of sediments and their surficial pathways from the mouth of Neretva River, through the Neretva Channel, toward the Adriatic Sea. This research was based on twelve box-cores and five grab samples collected within the Neretva Channel. Sediment dynamics were evaluated using several proxies, such as organic matter, radiochemical isotopes and select metal concentrations and physical parameters. The data analysis showed that the influence of the river on particle distribution along the Neretva Channel decreases northward, with an estimated sediment accumulation rate ranging from 1.9 to 8.5 mm/yr. The lowest accumulation rate was found in the sector not influenced by river inflow, whereas the preferential sediment accumulation area is in the center of the basin. We speculate that dispersion and accumulation of sediments are both driven by an eddy in the waters of the Neretva Channel triggered/or intensified seasonally by the interaction of karstic springs, river input and Adriatic Sea waters. Our results indicate that the anthropogenic factor does not affect the concentration of metals within the channel and that the river particles dynamics determine the Pb areal distribution, while Cr and Ni have a possible source located to the northwest of the river-mouth.

**Keywords:** Adriatic Sea; Neretva Channel; sediment dynamics; age model; metal concentrations

#### **1. Introduction**

The legal instruments of EU's Environmental policies and their regional and local application levels provide an interrelated regulatory framework for protection, preservation and prevention of the European oceans [1–4]. According to these European directives, they are intended to apply an ecosystem-based approach to manage human activities whilst ensuring sustainable use of marine goods and services (respectively) [4,5]. In particular, the Marine Strategy Framework Directive (MSFD) [3] uses eleven descriptors with several indicators covering ecological, physical, chemical and anthropogenic components of the ecosystems that need to be integrated to achieve a good environmental status [6]. In this context, the study of sediment dynamics plays an important role, as it provides the basic knowledge for a correct evaluation of the environmental quality.

The Adriatic Sea is an important sub-region of the Mediterranean marine area and has been proposed as Ecologically or Biologically Significant Areas in the Mediterranean (EBSAs) from the Barcelona Convention, UNEP/MAP [7]. Several studies were carried out on sediment dynamics and geochemistry in the Adriatic Sea, but they were mainly focused on its western side (Italian Economic Water Zone; see [8–14] and reference therein). Hence, few geochemical data are available from sedimentary deposits of the eastern coast [15–17] and potential input of contaminants is still little known.

Regional geology largely controls the composition of marine sediments; however, in densely populated regions, such as coastal areas, the anthropogenic influence may strongly affect the dispersion and concentration of organic matter and contaminants, modifying the river discharge through a combination of factors such as urban settlements and roads, runoff of agricultural soils, and dry and wet atmospheric deposition [18]. In this work, the influence of the Neretva River (NR) on the southern Croatian Adriatic coastal system was investigated in order to gain information on sediment dynamics of the area (Figure 1).

**Figure 1.** Study area and sampling locations. Shaded relief image derived from bathymetric data with sun illumination from NW, 45◦ over the horizon and no vertical exaggeration (grid resolution 10 m). Elevation and bathymetry data from EMODnet gridded data (www.emodnet.eu/bathymetry). Filled red and pink circles indicate box corer and grab samples locations, respectively. Bathymetry contour interval is 10 m. Red box in the inset indicates the study area.

The major aim of this work was to determine sediment pathways and dynamics of particles that bond metals in the Neretva Channel (NC) by combining sediment geochemistry, geochronological data and physical parameters of the water column to obtain basic information required for an assessment of the health and quality of the environment to preserve the marine life [19]. These results may provide a useful support for evaluation and decision-making processes aiming to achieve the environmental goals proposed by the Water Framework Directive [2] and the MSFD [3].

#### **2. Study Area**

The NC (Figure 1) is a narrow, semi-enclosed basin located along the southernmost part of the Croatian coast. Oriented SE–NW, it is bounded to the NE by the Croatian coastline, to the NW by the open Adriatic Sea through the Korcula Channel, and to the SW by the Peljesac Peninsula. The NC is a microtidal, low-wave energy environment with river-dominated sedimentation processes [15,17]. The river has an average annual water inflow of 332 m3/s with peaks in December and April, and with a long-term seasonal variability in terms of minimum and maximum monthly discharge from low to moderate, respectively [20]. The NR mouth, located ≈20 km to the east of the town of Metkovi´c, is characterized by a reclaimed alluvial plain and forms the largest depositional system in the southern Croatian coast, covering an area of <sup>≈</sup>246 km2.

The NR, despite high annual precipitation in the catchment area, has a low stream density because part of its water is collected into karstic aquifers [20]. Furthermore, five hydropower plants in the Neretva catchment impound a total area of 36 km2, storing a 1.1 km3 water reserve [21]. As a consequence, the present-day delta is not associated with the large volumes of deposits that other river systems along the western Adriatic coast are. Finally, water regime in the lower course of the NR is complicated by the interaction with seawater ingression. For instance, deepening of the NR riverbed may cause higher salinity of both water and soil due to a drop in the ground water level [17,22]. Salt water constantly penetrates into underground waterways, leading to the salinity of the soil, particularly during the dry season when the river flow is reduced [23]. Hence, the natural environment of the lower river course may be strongly threatened in terms of chemical, physical and biological changes [17], especially where human activities cause pressure, such as road construction sites, urbanization, hunting and mining [22,24,25]. Moreover, the NR is moderately affected by untreated municipal and industrial (metallurgy and other lighter activities) wastewaters.

#### **3. Materials and Methods**

During the oceanographic cruise NERES06 onboard the R/V Bios DVA in May 2006, twelve box cores (BC) and five grab samples (G) were collected in the study area (Figure 1). Sampling sites were selected along three transects parallel to the NW–SE-oriented axis of the channel in order to obtain an exhaustive coverage of the investigated area.

The BC short cores (≈20 cm long) were radiographed via a directional X-ray tube (M60 Gilardoni) using an aluminum and PVC filter, allowing the identification of the sedimentary structures before sub-sampling. Once the cores were opened, sediments were described for visual characteristics, and then, one-half of each was stored for the historical archive and the other half was sub-sampled with a frequency of 2 cm for sediment grain size and tracer determinations. Before analyses, sediments were dried at 60 ◦C in order to calculate the sediment porosity assuming a particle density of 2.5 g cm−<sup>3</sup> [26].

Total carbon (TC) and total nitrogen (TN) contents, and the stable isotopic composition, were determined on the surficial samples by using a FINNIGAN Delta Plus mass spectrometer, directly coupled to the FISONS NA2000 EA (for further details see [27]). The total organic carbon (TOC) was measured by a pre-treatment with 1.5 M HCl to remove the carbonates. TOC contents are reported as weight percent (wt%) on dry weight. Moreover, C/N was calculated as molar ratio between TOC% and TN%.

Grain size analysis was carried out on separates following wet sieving on a 63 μm mesh-size sieve in order to separate sand from finer fractions, after a pre-treatment with H2O2 to remove the organic matter and to favor disaggregation between sediment particles. Silt and clay fractions were determined with an X-ray Sedigraph Malvern Mastersizer 2000s.

The time framework was based on 137Cs and 210Pb radionuclides. 137Cs activity was measured by non-destructive gamma spectrometry (see [28] and references therein) using coaxial intrinsic germanium detectors (Ortec HPGe GMX-20195P and GEM-20200), while 210Pb was determined by chemical extraction of its daughter 210Pb, assuming secular equilibrium between the two isotopes [11,29].

In order to determine the metal fraction of surficial sediment particles and/or dissolved in the interstitial water, 0.5 g of de-frozen wet sediment was leached with HNO3 and H2O2 (10:3) under reflux [30].

Cr, Ni and Pb concentrations were determined by furnace atomic absorption spectrophotometers and the results were normalized to the sediment dry weight. Accuracy and precision were tested through repeated analysis of certified reference material NIST 2709, and in comparison with the reported values for the determination of labile or extractable elements. Results fell within the range of certified values. Accuracies (as %RDS), estimated by replicate analyses of NIST 2709, were 4% (for Ni and Cr) and 3% (for Pb). The leaching recoveries, calculated on the basis of certified values for total concentrations, were 66%, 63% and 93% for Pb, Cr and Ni, respectively.

Conductivity and temperature (CTD) data collections were carried out during the oceanographic cruise by correspondence with sampling stations with a Sea-Bird SBE 25 CTD equipped with temperature and conductivity sensors. The CTD data were processed according to UNESCO standards [31], and pressure values were averaged at 0.5 dbar intervals. The Ocean Data View software was used to interpolate spatially the CTD vertical profiles [32].

The areal distributions of metal contents, sand fraction and sediment accumulation rate (SAR) were computed by universal kriging interpolation, a method for which the values are modelled by a Gaussian process governed by prior covariance. Under suitable assumptions on the spatial continuity of the variable to interpolate, kriging gives the best linear unbiased prediction of the interpolated values [33]. Spatial continuity parameters (range, nugget and sill) were evaluated using theoretical models fitting the experimental variograms calculated for each variable [34,35]. Universal kriging assumes that a continuous property called "regionalized variable" consists of two parts: a drift, or expected value, and a residual, or deviation from the drift. The drift is modelled by a polynomial function within a given neighborhood. The residual surface obtained by drift removal can be regarded as first-order stationary in a statistical sense. However, the effectiveness of this technique depends on the correct specification of several parameters that describe the semivariogram and the model of drift. We assumed a linear model of the semivariogram implying that estimation error increases without bounds with increasing distance from the control point.

Spatial analysis and mapping were performed using the PLOTMAP [36] and the GMT [37] software packages adopting the WGS84 datum and a Mercator projection with standard parallel at the latitude of 42.5◦ N.

Data relationships were statistically analyzed through the Pearson correlation coefficient, using the STATISTICA software package.

#### **4. Results**

#### *4.1. Sediment Features*

Sediment grain size analyses indicate that the NC deposits are mostly clayey silts with a silty dominant fraction and a low-fraction of sand. In particular, on surficial samples, the coarser materials were found in the southernmost zone (sand content of 31% in site 15BC) and in the north-western sector of the channel (sand fractions of 51%, 39% and 31% at sites 07BC, 06BC and 10BC, respectively; Figure 2a).

**Figure 2.** (**a**) Log-normal sediment mean diameter [38] areal distribution and (**b**) organic carbon surficial areal distribution in the Neretva Channel with (**c**) covariation of N/C versus δ13C.

Surficial TOC (Figure 2b) varies from 0.6% (06BC) to 1.18% (17BC) and δ13C from -22.3‰ (06BC) to -27.0‰ (17BC). Total nitrogen (TN) content ranges from 0.15% (17BC) to 0.09% (06BC), resulting in C/N ratios from 5.9 to 8.5.

X-radiograph images, together with depth distributions of sediment porosity and sand fractions of selected BC samples, are shown in Figure 3. X-ray images of BCs do not display evident inner sedimentary structures; instead, bioturbations are observed in several sediment intervals; in particular, in cores from the area close to the river mouth (i.e. site 01BC). Sample porosities of the surficial level along the entire channel range between 0.62 and 0.71. In addition, at each sampling site porosity slightly decreases with depth (toward the bottom of the BC) without abrupt changes through the sedimentary sequence (Figure 3).

**Figure 3.** Examples of X-ray images, porosity and sand content depth profiles (left panel), and 210Pb and 137Cs concentration activity variations with depth (right panel) from selected box cores (BCs): (**a**) 01BC; (**b**) 03BC; (**c**) 05BC; (**d**) 06BC; (**e**) 10BC; (**f**) 14BC and (**g**) 15BC.

Vegetal remains are found mostly in sediments near the river mouth (sites 01BC, 02BC, 09BC and 15BC), whereas fragments of molluscs, bryozoans and ostracods are widespread in the surficial sediments of the NC, with the exception of the nearest site to the river mouth (01BC), where only fragments of molluscs are found, and of the deepest site (07G) where the benthic macrofauna is almost absent.

#### *4.2. Activity-Depth Profile of 210Pb and 137Cs Radionuclides*

137Cs and 210Pb content variations with depth of selected BC are shown in Figure 3. 137Cs measured only in three BC (Figure 3a–d), displays maximum activity concentration at the top of BCs (i.e. 01BC and 03BC or at sub-surficial levels as in 02BC and 05BC; Figure 3a,c) with values ranging from the detection limit to 18 Bq/kg (Figure 3d,e). Unfortunately, the 1986 peak of 137Cs relative to the Chernobyl accident cannot be resolved in our samples.

210Pb contents in cores generally decrease with depth following a regular exponential trend (Figure 3). Since cores 04BC, 09BC, 13BC, 14BC and 15BC are strongly affected by physical or biological mixing (see X-radiograph; i.e. Figure 3f,g), a different 210Pb pattern is observed.

#### *4.3. Regional Water Circulation and Water-Column Structure*

During the sampling survey, temperature (T) and salinity (S) values in the water column ranged from 12.7 to 17.1 ◦C and from 19.2 to 38.0 psu, respectively (Figure 4a,b). At sites 01BC and 17BC, in front of the river mouth, recorded values of T and S at sea surface were ≈ 16 and 15.2 ◦C, and ≈ 34 and 19.3 psu respectively, whereas values at the sea bottom displayed lower temperature and higher salinity (≈ 13 and 13.5 ◦C, and ≈ 37.5 and 37.6 psu respectively).

**Figure 4.** (**a**) Temperature and (**b**) salinity spatial distribution. Temperature (**e**,**f**) and salinity (**g**,**h**) vertical cross sections along the northern (**c**) and the central (**d**) transects running parallel to the NE coast of the Neretva Channel.

T and S profiles across the water column are shown in Figure 4e–h. Observed values reveal water stratification with the highest temperatures at the surface in the southeast sector of the NC. The northern transect (Figure 4e,f) and surficial areal data (Figure 4a,b) suggest water inflow with lower temperature and salinity localized along the coast to the north of the main Neretva river-mouth

in the vicinity of station 11G, where a local decrease of salinity has been also recorded at 0 and 5 m water-depth intervals (Figure 4e,h). The water mass with lower T and S moves deeper toward the center of the channel.

#### *4.4. Metal Distributions*

Among the several metals, we focused on Cr, Ni and Pb in this study, because their origin may be due to both natural and anthropogenic causes. The spatial distributions of their concentrations (μg/g) allowed identifying zones of provenance (source areas) and areas at critical levels (Figure 5). Results shown in Figure 5 indicate that site 08G, in the northern part of the investigated area, displays the maximum contents of Ni and Cr (84.0 and 71.5 μg/g respectively), whereas the minimum contents were recorded at site 06BC (28.6 and 28.7 μg/g, respectively) located in center of the northern sector of the NC (Figure 5b,c). In particular, surficial Cr and Ni concentrations decrease toward the river mouth. Differently, the Pb concentrations show two different areas of local maxima: the first (36 μg/g) is located in front of the river mouth at site 01BC; the second (67.9 μg/g) in the center of the NC at site 10BC.

**Figure 5.** Surficial distribution of (**a**) Pb, (**b**) Ni and (**c**) Cr concentrations.

#### **5. Discussions**

#### *5.1. Surface Sediment Distribution*

The Neretva's surficial sediment is generally composed of clayey silt. The distribution suggests an influence of the river plume decreasing along the SE–NW direction (Figure 2a). In particular, accumulation of fine-grained particles (mean grain size > 7.5 ϕ [38]) occurs mainly in two areas: in front of the NR mouth and in the central/northeastern part of the NC (Figure 2). This accumulation pattern is in agreement with the observed water circulation, where a hypopycnal river plume formed at the mouth distributes fine-grained particles over the entire channel area [16,17]. These observations suggest a different nature of the two observed coarser-grains depositional zones. The southernmost

zone may only receive river sediments sporadically during events of strong hydro-dynamism in the channel and may be considered a non-depositional area characterized by relict coarser sediment. This silt-sandy accumulation zone was previously reported by [16]. Instead, the northernmost sandy zone may be related to the influence of high energy currents from the open Adriatic Sea that remove the finest part of the sediments.

The observed porosities and particle grain size in the BCs (Figure 3) suggest that over the time recorded by our samples, the NC sedimentation was continuous in a low-energy environment. Moreover, the chaotic textures of sediments in the cores nearest to the river pro-delta area (BC01-17 and 13) suggest also that these deposits are affected by several post-depositional processes, including physical mixing and biological reworking due to microbial activity. Both of those processes may be related to the seasonal river input variability or episodic flooding events from the NR.

#### *5.2. Organic Carbon*

Relationships between C/N ratios and the δ13C contents are commonly used as proxies to determine the source of organic matter in coastal areas [26,39,40]. In addition, the δ13C and δ15N isotopic compositions and the N/C ratio help to discriminate the marine versus terrestrial origins of the particles [41]. Since the photosynthetic processes and carbon sources are different between marine organisms and terrestrial plants, there is an inverse correlation between the δ13C and C/N ratio: A high δ13C values suggests a predominant influence of marine phytoplankton [42]. On the contrary, low δ13C values (lower than -27‰) are indicative of vegetable sources (C3 plants) from inland; i.e., proximity to the shoreline and/or proximity to a source of organic matter of continental origin [43]. Here, we considered the N/C ratio instead of the C/N, because it behaves linearly in a mixing model [43,44] and an estimate of the sedimentary organic carbon fraction is more reliable [45]. The observed N/C and δ13C distributions (Figure 2c) indicate that the sites close to the river mouth (01BC and 17BC) have a clear terrestrial signature due to the strong fluvial input. The marine input increases northward, and the transition between terrestrial and marine regimes may be located along the theoretical line joining sites 12G, 4BC and 9BC (Figure 2b). In addition, a strong negative correlation (−0.81) is observed between TOC and <sup>δ</sup>13C, a result in agreement with that of [17]. These data together with a lack of correlation with the clay fraction (Table 1), suggest riverine 13C depleted terrestrial organic matter as the main source of sedimentary organic carbon [27].

**Table 1.** Linear correlation coefficients between Pb, Cr and Ni concentrations (μg/g), porosity, organic carbon (TOC, %), total carbon (TC, %), inorganic carbon (TIC, %), C/N ratio, carbon stable isotope (δ13C, ‰) and sediment compositions (fines, sand, clay, silt contents, %). Statistically relevant coefficients are reported in bold and highly significant one in bold red.


These results allow us to hypothesize that the input material from the NR is quickly moved northward by marine currents and released not far from the river mouth, particularly in the area near and along the coast (north of the river mouth). The observed organic carbon distribution highlights a reduced fluvial influence moving seaward. However, surficial sediments may be affected by seasonal conditions and their distribution may be different in other periods of the year due to different amounts of input material from the river.

#### *5.3. Age Model and Sediment Accumulation Rates*

Activity-depth profiles of 137Cs and 210Pb were evaluated along the BCs to define an age model of the NC recent sedimentary sequence (Figure 3).

Correlations between 210Pb age determinations and depth are strongly model-dependent in cores with non-exponential 210Pb profiles [46]. Assuming a constant 210Pb flux below the surficial mixed layer, a "constant flux–constant sedimentation" model has been adopted to obtain an estimate of sedimentation accumulation rates (SARs) for the last hundred years. However, since the main assumptions of most common conceptual models in sediment chronologies refer to the inputs of particles and/or a radiotracer onto the sediment, in those samples where 210Pb profiles were not suitable for model calculations (such as in samples 06BC and 10BC), rough estimates of sediment accumulation rates were obtained by assuming an age of 100 years at the depth where the 210Pb background value of ≈18 Bq/kg is found (Figure 3d,e). Since physical mixing and bioturbation were neglected in the calculations, these apparent average rates have to be considered upper limits.

Since the 210Pb dating model requires validation by a second, independent stratigraphic tracer [47,48], 137Cs activity-depth profiles were considered in the calculations. Unfortunately, as mentioned above, the 1986 137Cs peak (Chernobyl accident) is not cleatracer [47,48]rly identifiable in the investigated sedimentary intervals. This is probably due to the strong bioturbation and sediment mixing (see X-ray images in Figure 3f,g) as suggested before by [16]. Moreover, since the half-life of 137Cs is 30.17 years, it could not be excluded that the concentration is below the experimental sensitivity. This effect may be enhanced in case of high contents of carbonate (such as in karstic areas) that are less reactive to radiotracers [49]. In order to overcome validation troubleshooting, the core depth where 137Cs reached the instrumental detection limit was used as reference age, assuming it corresponded to the early 1960s, when a large number of nuclear bombs were exploded in the atmosphere.

Estimated sediment accumulation rates (SARs; Figure 6) range from 1.9 to 8.5 mm/yr. These values differ partially from those previously estimated by other authors [16], ranging from 4 to 6 mm/yr and based only on the distribution of 137Cs in core sediments. In particular, the lowest SAR value of 1.9 mm/yr was found at site 06BC (Figure 6), located to the NW of the channel at the boundary with the Korcula Channel, a location comparable to that of sample K3 of [16] where a SAR of 4 mm/yr was estimated. In the sectors of the channel less influenced by the river outflow, SARs range from 1.9 to 2.3 mm/yr, while in the areas where the river influence is high, SARs range from 4.0 to 8.5 mm/yr. The highest accumulation rates are observed in front of the river mouth and along the northeastern coast with a peak of localized preferential accumulation at site 04BC (8.5 mm/yr; Figure 6). Estimated accumulation rates are in agreement with the observed water circulation, where a hysopycnal river plume formed at the river mouth distributes particles toward the open sea.

**Figure 6.** Apparent sediment accumulation rates (SAR) over the study area.

Apparent accumulation rates (Figures 3 and 6) suggest that most surficial sediments in the NC result from the present day seasonally river-dominated coastal sedimentary regime, with few relict deposits in its southernmost sector, south of the river mouth. According to the calculated SARs, our BCs record a time interval ranging from 18 (04BC) to 100 years (06BC) ago.

#### *5.4. Regional Water Circulation and Water-Column Physical Properties*

Surficial currents along the eastern Adriatic in the proximity of the NC are always oriented from SE to NW following the overall cyclonic circulation of the basin [50]. Local data on annual mean values of water circulation within the NC are not available in the literature, except for a little punctual information. Given that the NC represents a semi-closed marine environment, it is reasonable to assume that channel hydrodynamics are the result of a balance between tidal currents and fluvial influence, and seasonal variations in precipitation may act as a driving factor for the regional water exchange with the Adriatic open sea.

Generally, river currents do not exceed the velocity of 0.1 m/s and are higher during the winter season [51,52]. In particular, in 2004, daily outflows (m3/s) during winter and spring periods ranged from 300 to 400 m3/s, with short time intervals of peak flow (>500 m3/s) or reduced flow (< 200 m3/s). However, during the summer season, the river outflow is drastically reduced (up to 50 m3/s), as reported by UNEP/MAP [53]. In addition, horizontal distribution of freshwater into the channel may depend on the direction of local winds and may vary on a daily basis [54].

Based on CTD data, sediment grain size and SAR surficial distributions, we suggest the presence of an eddy in the NC waters located at the center of the channel that may strongly influence sediment dispersion and determine areas of localized preferential sediment accumulation not necessarily in correspondence with the river mouth (e.g., 10BC, Figure 4). Moreover, a considerable amount of freshwater is discharged into the channel by several submarine karstic springs [55], suggesting high inflow of cold water that lacks sediment load during periods of intense rainfall.

The observed water stratification along the northern transect may represent a seasonal snapshot of the hydrologic conditions of the study area. Mixed-water conditions may dominate through the entire NC area during flooding events in autumn. However, for the objectives of this study, spring conditions may be considered the optimal hydrological situation. Close to the river mouth, less salty and colder water was detected. This could be due to a reduction of the salt-wedge during low tide conditions and/or intense river outflow, or the karstic characteristics of the region. Karstic fresh waters have generally poor or null suspended material contents, explaining why these inflows do not affect organic matter or metal distributions.

#### *5.5. Metal Distributions*

Metal concentrations suggest that the river outflow plays only a marginal role in Pb, Cr and Ni surficial distributions (Figure 5). In detail, Pb shows the highest concentration at site 10BC located in the center of the channel and far from possible direct sources. This may be due to an event of local discharge or to a pulse in the transport of metal enriched sediment, as suggested by the higher sand content at the top of the core and by the N/C values (Figure 2b,c and Figure 3). A second local maximum of Pb content is located in front of the river mouth following a distribution inversely correlated with the distance from the mouth (Figure 5a).

Cr and Ni contents (Figure 5b,c) are affected by a source located along the coast to the NW of the river mouth (close to station 08G). Moreover, Cr content shows another local maximum at site 07G (70.4 μg/g), where an increase in the sand fraction is observed, suggesting an additional source. On the other hand, the minimum values of Ni and Cr contents at site 06BC may indicate a scouring resuspension or not sedimentation, as suggested by the observed N/C values. In the southern part of the NC, where coarser sediments are located, an increase of Cr and Ni contents associated to a low organic carbon and to a low sediment accumulation rate of 2.3 mm/yr (15BC; Figure 3) suggests that this zone acts as a trap for metal bearing particles. Alternatively, the Cr and Ni abundances in this area may be influenced by sediment inputs from Peljesac Peninsula.

Some authors indicate the ratio Cr/Ni as a possible tracer of geo-genic versus anthropogenic influences [56–58] and suggest it as suitable to determining geochemical baselines in the case of high natural concentrations [34,35,59]. In particular, Cr concentrations in south Dalmatia show a mean value of 126 μg/g; that is the highest value found in karstic regions, probably due to the presence of chromite-bearing ultramafic rocks and/or of clastic deposits derived from older mafic magmatic rocks. Similarly, Ni levels in the same area show a mean value as high as 84 μg/g as a consequence of the presence of mafic and ultramafic rock outcroppings in the region. These rocks are probably a source of Ni in alluvial soils of the floodplains of rivers draining this region [57]. The expected Cr/Ni value reported in literature [56–58] is about 1.5, whereas in the study area Cr/Ni ranges from 0.7 (site 12G) to 1.9 (site 07G) with a mean value of 0.9. This discrepancy may be due to a natural variability but also due to the extraction efficiency of the leaching method. However, ratio values are generally uniform in the area supporting the hypothesis of Cr and Ni background concentrations.

According to previous data [17], heavy metals (such as Pb, Cr and Ni) in surficial sediments are tightly linked to fine grain size fractions, and in particular, to clay minerals coated by iron oxides/oxy-hydroxides that may trap metal cations within their crystal-chemical structures. Univariate statistical analysis shown in Table 1 confirms the previous observations for Ni and Cr: that they correlate positively with each other, and with clay and porosity, while they correlate negatively with total and inorganic carbon. In addition, organic carbon shows a positive correlation with fine grain size, in particular, with the clay fraction, and negative ones with total carbon and inorganic carbon. These results may support the hypothesis of clay minerals and organic matter as driving factors of Cr and Ni areal distribution. Pb shows a different behavior, characterized by a negative significant correlation with porosity. Since porosity is correlated negatively to silt and positively to clay, this suggests that Pb is not immobilized in the clay fraction of the sediments having higher affinity with the silt fraction.

A preliminary evaluation of environmental risks due to contaminant loads within the channel has been carried out, comparing surficial concentration values with quality criteria for marine sediments (Table 2)—threshold effect level (TEL), effect range low (ERL), probable effect level (PEL) and effect range median (ERM) [60–62]—and with benchmarks (LCB and LCL) for Italian coastal areas [63]. These guidelines are screening tools to predict potential sediment toxicity, linking sediment metal concentrations to the adverse biological effects that may result from exposure to chemicals.


**Table 2.** Minimum, maximum and average concentrations (in μg/g d.w.) of heavy metals in box-core sediments. Comparison with Italian and international sediment quality guidelines benchmarks.

Mean metal concentrations pre-seventies (BG); world average (WA) [64]; \* Italian Sediment Quality benchmark: APAT and ICRAM [63]; \*\* International Sediment Quality Guidelines [60–62].

The Pb and Cr levels in surficial samples are below the benchmarks of PEL, ERM, and the Italian LCL guidelines, except at sites 10BC and 14BC for Pb (ERL and LCB) and 08G for Cr (ERL). However, the incidence of the effect for these samples may be considered low or scarcely probable (8%–30% for Pb and 2.9%–21.1% for Cr; [62–64]).

The Ni surficial concentration exceeds, at several sites, both PEL and ERM guidelines, whereas stations 06BC and 07G and 10BC comply under PEL; however, the Italian benchmark LCB is exceeded only at site 08G. In general, the effect incidence for exceeding ERM is common, although for Ni it is only 16.9% [62–64]. The Pb, Cr and Ni concentrations in surficial sediments of the Neretva Channel fall into an unpolluted class with exception of Pb content of sample 10BC falling into a moderately polluted class. All this highlights a general unpolluted situation of the channel.

The sediment flux (carried out where BCs were available; Figure 7) provides a quantitative estimation of the amount of material that deposits in the channel. Two areas of higher flux are identified: in front the river mouth and in the center of the NC, at site 04BC. The latter location is not related to any particular morphologic setting of the channel and coincides with the area where we suggest the presence of an eddy in the water circulation, probably favoring a critical area of accumulation. TOC, following a similar distribution, tells the same story (Figure 2b), given that TOC input depends mostly from the river inflow because springs water may be assumed containing very few particles. These observations document that local hydrology plays a key role in controlling the sediment distribution in the area.

**Figure 7.** Areal distribution of sediment flux (g/cm<sup>2</sup> yr) in surficial sediment of the Neretva Channel.

#### **6. Conclusions**

In this study, a multi-proxy approach on surficial sediment samples allowed us to determine sediment dynamics of the Neretva delta system, the largest modern siliciclastic depositional system on the eastern Adriatic coast.

Samples from 12 box-corer and five grab stations document that surficial sediments are predominantly clayey silt and coarse silt with a distribution of finer grained particles decreasing southward moving away from the river mouth.

Estimated accumulation rates, based on 210Pb and 137Cs activity versus depth profiles, range from 1.9 to 8.4 mm/yr, suggesting the center of the channel and the river mouth as the sectors with the highest sediment deposition rates. In particular, two areas of higher sediment accumulation were identified: (i) close to the river mouth; and (ii) in the central-northern sector of the Neretva Channel along the axis of the basin.

Temperature and salinity vertical distributions indicate stratified waters along the axis of the basin close to the Croatian coast, and the presence of less salty and colder water in the proximity of the river mouth. The latter consideration, together with the observation of strong salt intrusions in the NR delta, suggests a salt wedge reduction during low tide conditions and submerged fresh water inputs. Moreover, relationships between TOC and TN and their spatial distribution, in particular, covariations between N/C and δ13C, indicate a terrestrial origin with prevalent accumulation in the vicinity of the river mouth that decreases northward towards the open sea. The scenario depicted above agrees with temperature and salinity data of the water column and their surficial distribution in the channel.

Finally, based on CTD data, grain size, TOC and SAR distributions, we suggest the presence of an eddy in the NC waters near the center of the channel that may be triggered and/or intensified seasonally. This may strongly affect sediment dispersion favoring areas of localized preferential sediment accumulation not necessarily close to the river mouth (e.g., 10BC, Figure 4).

The distribution of Pb contents in the surficial sediments follows a general trend inversely proportional to the distance of the river mouth with two local zones of accumulation: in front of the river mouth and in the center of the channel. Univariate statistics indicate that Pb is not immobilized in the clay fraction. On the contrary, Ni and Cr distributions indicate that they follow the clay fraction and that they account for other sources besides the river, such as sediment input from the Peljesac peninsula. The comparison with marine sediment guidelines shows that surficial concentrations of the examined metals have a low toxicity.

**Author Contributions:** F.G. had the role of project administrator and investigator in this study. S.R. wrote the preliminary version of the manuscript with the support of F.G. and the other authors. The methodology and validation of data were carried out by S.A., L.C. and M.L. F.C. reviewed and edited text and figures, and curated data. All authors discussed the results and contributed to the final manuscript. M.R. supervised the scientific project and provided financial support for the scientific work and publication. All authors have read and agreed to the published version of the manuscript.

**Funding:** Funds were provided, within the framework of ADRICOSM-NERES project (2006–2007) (Environmental regeneration and sustainable development of the delta of the NEretva river), by the Italian Ministry of Foreign Affairs and the Ecoss project Italy-Croatia (2019).

**Acknowledgments:** Thanks to Frano Matic, Grozdan Kuspilic and Slavica Matijevc of the Institute of Oceanography and Fisheries, Split Croatia; Gabriele Marozzi and Costante Luttazzi, for sample collection, sampling and treatment on R/V Bios; Mauro Frignani for helpful suggestions; and Nadia Pinardi for project coordination. Special thanks to Prof. Enrico Bonatti of Earth Institute Lamont-Doherty Earth Observatory, Columbia University for the careful english language revision of the manuscript. This is contribution number 2022 of the Istituto Di Scienze Marine of the Consiglio Nazionale delle Ricerche, Bologna.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study, in collection, analyses, or interpretation of the data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Impact of Dredged Material Disposal on Heavy Metal Concentrations and Benthic Communities in Huangmao Island Marine Dumping Area near Pearl River Estuary**

**Wei Tao 1,†, Zhongchen Jiang 1,†, Xiaojuan Peng 1, Zhenxiong Yang 1, Weixu Cai 1, Huili Yu <sup>2</sup> and Jianjun Ye 1,\***


**\*** Correspondence: jjye@live.com; Tel.: +86-134-8025-9901

† These authors contributed equally to this work.

**Abstract:** The Huangmao Island dumping area is adjacent to the Pearl River Estuary in the South China Sea. From its first dumping activity in 1986 to 2017, 6750 <sup>×</sup> <sup>10</sup><sup>4</sup> <sup>m</sup><sup>3</sup> dredged materials were dumped in this dumping area. Sediment pollution levels, ecological risk, and benthic communities in 2011–2017 were evaluated; the results showed that the concentrations of the heavy metals (HMs; except Hg) in surface sediments of the dumping area met the class I standard of marine sediment quality (GB 18668-2002). HMs in the surface sediments were relatively high in the northern and central areas but relatively low in the south of the dumping area. Speculation was that the spatial variation in HM concentrations might be caused by dumping activities. The Nemerow index implied that the contaminated area was mainly in the north of the dumping area (S1, S2, and S3), where the dumping amount was the largest. The potential ecological risk (*E<sup>i</sup> r*) indices of Zn, As, Cu, and Pb indicate that these metals posed a low risk to the ecosystem of the dumping area, whereas Cd and Hg posed a high risk at some stations. The geoaccumulation indices (*Igeo*) of Zn, As, Cu, and Pb specified no pollution or light pollution in the study area, whereas those of Cd and Hg in most years indicated mild contamination levels. Benthic organisms in the study area were arthropods, chordates, annelids, mollusks, echinoderms, nemertinean, coelenterate, and echiuran, among which arthropods were the most abundant. The abundance of taxa and density of benthic organisms had a little difference among the stations within the dumping area, but were significantly lower than those of the stations outside the dumping area. In addition, non-metric multidimensional scaling analysis confirmed that the observed patterns separated the stations within the dumping area from stations outside the dumping area. The evaluation results of the HMs revealed that the dumping area with a large dumping amount was more severely polluted. Dumping dredged materials seemed to have a negative impact on the benthic community in the dumping area.

**Keywords:** marine dumping area; heavy metal; dredged material; benthic community; Huangmao Island

#### **1. Introduction**

In urbanization and the rapid development of the marine transportation industry and coastal engineering construction projects, many dredged materials are produced [1,2]. Many disposal methods are available for dredged materials, but land disposal is a priority, such as reclaiming land from the sea and making solidified materials [3]. For cost and complexity reasons, another common disposal method for dredged materials is marine dumping [4]. From 2011 to 2017, the volume of dumping materials in China was estimated to be 11.2 × <sup>10</sup><sup>8</sup> m3 [5]. Dredging and disposal are serious environmental concerns in coastal management [6,7].

Dredged materials mainly originate from harbors and channels. These areas are characterized by low hydrodynamic force, poor automatic purification capacity, low dissolved

**Citation:** Tao, W.; Jiang, Z.; Peng, X.; Yang, Z.; Cai, W.; Yu, H.; Ye, J. Impact of Dredged Material Disposal on Heavy Metal Concentrations and Benthic Communities in Huangmao Island Marine Dumping Area near Pearl River Estuary. *Appl. Sci.* **2021**, *11*, 9412. https://doi.org/10.3390/ app11209412

Academic Editor: Mauro Marini

Received: 3 September 2021 Accepted: 29 September 2021 Published: 11 October 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

oxygen levels, intensive human activities, and high HM contents [8]. HMs cannot be biodegraded, accumulate rapidly, and reach a toxic level within a short time. During dumping, these toxic, persistent, and bioaccumulative HMs enter sediments through the decomposition, sedimentation, absorption, and formation of complexes and eventually harm benthic ecosystems [9,10]. Meanwhile, their removal is difficult and sometimes impossible [11,12]. Therefore, sediments are usually considered indicators of HM pollution [13].

Dumping dredged materials can cause the resuspension of marine sediments [14], resulting in the release of HMs from sediments into overlying seawater. In addition, this practice leads to an increase in seawater turbidity and a decrease in seawater depth, which interferes with the respiration and feeding of marine organisms and affects the regional hydrodynamic force [3]. Studies have indicated that dumping can have various impacts, ranging from obvious and long-term [4,15] to unobvious and short-term damage in the benthic community [16,17]. Harvey et al. [4] and Roberts et al. [18] suggested that the diversity of communities in the dumping area might be reduced, dominant patterns within the community might be altered, the abundance of some species might decrease, and the abundance of opportunistic species might increase.

The Huangmao Island marine dumping area is near the Pearl River Estuary in the South China Sea. The dumping area began to receive dumping materials in 1986. As of 2017, this dumping area had received 6750 × 104 m3 of dredged materials. Dumping materials in the Huangmao Island marine dumping area were mainly dredged materials from wharves, harbor basins, and channels: Most were cleaning dredged materials (Class I standard for dredged materials for dumping in dumping area) [19], a small amount was contaminated dredged materials (Class II), and none were polluted dredged materials (Class III).

To alleviate the marine environment during the use of the dumping area, monitoring and evaluating the dumping area is necessary, especially in terms of sediments and benthos. A few studies have been carried out to evaluate the effect of dumping activities to marine sediments and benthic community [3,17,20,21], but few studies have been conducted in the South China Sea, where there are 25 dumping areas as of 2017. Among these dumping areas, the Huangmao Island dumping area is the earliest to received dredged materials. Therefore, it is highly necessary to perform assessment of the dredged material disposal in this dumping area. Based on years of monitoring of the Huangmao Island marine dumping area, an evaluation of the pollution status of sediments in the Huangmao Island marine dumping area was carried out by the geoaccumulation index and ecological risk index. The effects of dumping on the benthic community were investigated using non-metric multidimensional scaling analysis (n-MDS). According to the research results, suggestions and measures are presented to provide technical support for the sustainable use of the dumping area.

#### **2. Materials and Methods**

#### *2.1. Description of the Study Area*

The Huangmao Island marine dumping area is adjacent to the Pearl River Estuary, southeast of Hengqin Island, south of Huangmao Island, and northwest of the Wanshan Islands (Figure 1). The geographical coordinates are 113◦38- 30--–113◦40- 30-- E and 21◦58- 00--–22◦01- 00-- N, and the dumping area is approximately 20.6 km2. The Huangmao Island marine dumping area was approved in 1986. The dumping volume in the early stage was small but increased significantly since 2011. From 2011 to 2017, a total volume of <sup>4762</sup> × 104 <sup>m</sup><sup>3</sup> of dredged materials was dumped in this area. The dumping volume was the largest (1607 × <sup>10</sup><sup>4</sup> m3) in 2016 and the smallest (134 × <sup>10</sup><sup>4</sup> m3) in 2017. The sediment type in the dumping area was mainly silt, followed by sand and clay. The water depth of the dumping area gradually increased from the northwest to the southeast. In 2011, the water depth of the dumping area ranged from 8.2 to 14.4 m, with an average of 12.2 m. In 2017, the water depth ranged from 9.4 to 14.5 m, and the average value decreased by 0.3 m, of which the maximum decrease (1.2 m) was in the northwest of the dumping area.

**Figure 1.** Sampling stations at the dumping area and adjacent sea areas.

#### *2.2. Sampling*

Surface sediment samples (0–5 cm) were collected from five sampling stations in October each year from 2011 to 2017. Because only a small amount of data was collected in 2015, the situation in 2015 was not included in this study. In addition, a benthic community investigation was conducted in the study area in 2017. Sediments were collected 2–3 times at each site by a grab sampler and then mixed and packed into pre-cleaned glass jars and frozen at −20 ◦C until further treatment. Benthos samples were collected using qualitative and quantitative methods. Quantitative samples of benthos were collected using a grab sampler. The samples were washed with seawater through a sieve with a diameter of 1 mm, and all biological samples were collected and transferred to a plastic container. Qualitative samples were collected with a 1.0 m wide Agassiz trawl. After dragging slowly for 10 min (approximately 200 m) at each station, all benthic organisms from the trawl were collected. Samples were bottled, numbered, and fixed with a 5% neutral formalin solution. Species were identified, counted, and weighed in the laboratory.

#### *2.3. Laboratory Analysis*

Before chemical analysis, sediments were freeze-dried to a constant weight. After removing the gravel and shells, the dried sediments were ground and sieved through a 96 μm stainless-steel sieve. Sediment samples (0.2000 g) for the measurement of Hg and As were digested with 10.0 mL of a mixture of acid (HNO3+HCl). Sediment samples (0.1000 g) for the measurement of Cu, Zn, Cd, and Pb were digested with a mixture of concentrated HNO3 (1.0 mL) and HClO4 (2.0 mL). HMs (Cu, Pb, Zn, and Cd) in the sediment samples were determined using a flame atomic absorption spectrometer (Analytik Jenna ContrAA 700). As and Hg in the sediment samples were tested using an atomic fluorescence spectrometer (Beijing Haiguang AFS 9560). Sulfide and TOC in sediment samples were determined by methylene blue spectrophotometry and potassium dichromate volumetric method, respectively. Oils were analyzed using a UV spectrophotometer (Shimadzu UV-2450). Eh and pH values of sediments were measured with a potentiometer and a pH meter, respectively. Although As is a metalloid that exhibits intermediate properties between those of metals and non-metals, this text refers to it as a metal. Organisms were sorted, counted, and identified to species level.

Quality assurance and quality control were evaluated using duplicates, blanks, and standard reference material (GB W07333) from the National Research Center for Standard of China. The detection limits of Hg, As, Cu, Pb, Cd, and Zn were 0.002, 0.06, 0.1, 0.1, 0.02, and 0.2 mg/Kg, respectively. All chemicals used for the analysis were of analytical grade or above. Blanks and duplicates were run for each batch of 10 samples. The blank values are below the detection limit. From the values of the duplicates, the relative errors of Hg, As, Cu, Pb, Cd, and Zn were below 6.5%, 1.8%, 2.5%, 3.7%, 5.9%, and 2.1%, respectively. The measured values of the standard reference material were within the error allowed. The results of quality assurance and quality control indicate that the accuracy and precision are acceptable.

#### *2.4. Pollution and Ecological Risk Assessment Methods*

2.4.1. Nemerow Pollution Index (*Pi*)

The Nemerow pollution index (*Pi*) was used to evaluate the sediments, as follows [22]:

$$P\_{ij} = \mathbb{C}\_{ij} / \mathbb{S}\_i$$

$$P\_{ij\text{ave}} = \frac{1}{\text{m}} \sum\_{i=1}^{\text{m}} P\_{ij}$$

$$P\_i = \left\{ \left[ \left( P\_{ij\text{max}} \right)^2 + \left( P\_{ij\text{ave}} \right)^2 \right] / 2 \right\}^{1/2}$$

where *Pij* is the standard index of the *i*th evaluation factor, *Cij* is the measured concentration of the *i*th evaluation factor, *Si* is the evaluation standard of the *i*th evaluation factor, m represents the number of evaluation factors, and *Pijave* and *Pijmax* refer to the average and maximum single factor pollution indices, respectively. The Nemerow pollution index was divided into five zones to describe pollution levels: unpolluted (≤0.7), lightly polluted (0.7 < *Pi* ≤ 1), mildly polluted (1 < *Pi*≤ 2), middle-level polluted (2 < *Pi* ≤ 3), and seriously polluted (*Pi* > 3).

#### 2.4.2. Geochemical Accumulation Index (*Igeo*)

The pollution status of HMs in sediments was evaluated using the geochemical accumulation index (*Igeo*) as follows [23]:

$$I\_{\rm geo} = \log\_2 \frac{c\_i}{1.5 \times c\_{Bi}}$$

where *Ci* is the measured concentration of element *i* in the sediment, and *CBi* refers to the geochemical background value of an element. The geochemical accumulation index is divided into six zones to describe pollution levels: clean (*Igeo* < 0), light pollution (0 ≤ *Igeo* < 1), mild contamination (1 ≤ *Igeo* < 2), middle-level pollution (2 ≤ *Igeo* < 3), and serious contamination (*Igeo* ≥ 3).

#### 2.4.3. Integrated Potential Ecological Risk Index (*RI*)

The ecological risk was evaluated using the potential ecological hazard index (*RI*) as follows [24]:

 $E\_r^i = T\_r^i \times C\_f^i = T\_r^i \times \frac{Ci}{C\_{Bi}}$   $RI = \sum\_{i=1}^n E\_r^i$ 

where *C<sup>i</sup> <sup>f</sup>* is the accumulation factor of metal *<sup>i</sup>*, expressed as *<sup>C</sup><sup>i</sup> <sup>f</sup>* <sup>=</sup> *Ci CBi* ; *Ci* is the concentration of metal *i* in the sample; *CBi* is the geochemical background value of metal *i* in the sediments; and *Tr <sup>i</sup>* is the toxicity coefficient of metal *i*. *Ei <sup>r</sup>* is the individual ecological risk of metal *i*, and *RI* represents the potential ecological risk caused by overall contamination, which is the sum of all risk coefficients for metals.

Based on *E<sup>i</sup> <sup>r</sup>* value, the potential risk is classified into five categories: low risk (*Ei <sup>r</sup>* ≤ 40), middle risk (40 < *<sup>E</sup><sup>i</sup> <sup>r</sup>* ≤ 80), relatively high risk (80 < *<sup>E</sup><sup>i</sup> <sup>r</sup>* ≤ 160), high risk (160 < *Ei <sup>r</sup>* ≤ 320), and extra-high risk (*Ei <sup>r</sup>* > 320). *RI* is classified into four categories to describe integrated potential ecological risk: low risk (*RI* < 150), middle risk (150 ≤ *RI* < 300), relatively high risk (300 ≤ *RI* < 600), and high risk (*RI* > 600).

#### *2.5. Statistical Analysis*

SPSS 25 was used to evaluate the correlation between HMs and environmental factors. Multivariate analysis was performed after the fourth root transformation of the abundance data from each sampling station. Outputs from n-MDS ordination models of the Bray– Curtis similarity matrix were obtained. For benthos data, n-MDS was conducted based

on the abundance of taxa, density, and diversity of each sampling station. Multivariate analysis was conducted using PRIMER V5 software [25].

#### **3. Results and Discussion**

#### *3.1. HMs in the Sediments*

The concentrations of Hg, As, Cu, Pb, Cd, and Zn in surface sediments of the dumping area were 0.030–0.26, 7.90–19.8, 13.1–36.3, 17.4–54.7, 0.11–0.34, and 50.0–122 mg/kg, respectively, with an average of 0.080, 15.1, 24.5, 30.8, 0.21, and 97.6 mg/kg, respectively (Table 1).


**Table 1.** The concentrations of HMs in the dumping area (unit: mg/kg).

<sup>a</sup> Average value of all the samples over six years. <sup>b</sup> CV%=average/standard deviation. <sup>c</sup> Standard of the Marine Sediment Quality Standard (MSQS) of China (GB 18668-2002). <sup>d</sup> Threshold effect level. <sup>e</sup> Probable effect level. <sup>f</sup> Effect range low from the National Oceanic and Atmospheric Administration (NOAA). <sup>j</sup> Effect range medium from the National Oceanic and Atmospheric Administration (NOAA).

> If the concentrations of pollutants in the sediments are lower than the threshold effect level (TEL), adverse biological effects are expected to rarely occur. If the concentrations of pollutants in the sediments are higher than the probable effect level (PEL), adverse effects are expected to frequently occur. If the concentrations of pollutants in the sediment are between the TEL and PEL, the probability of adverse biological effects is close to that of no adverse biological effects. In this study, the concentration of As in the sediments was higher than the TEL in each year, that of Cu was higher than the TEL except in 2013 and 2017, and that of Pb was lower than the TEL except in 2011 and 2012. The concentrations of Hg, Zn, and Cd were lower than the TEL. The concentrations of all HMs were lower than the PEL. The HMs in sediments from the dumping area may have adverse biological effects. The effect range low (ERL) and effect range medium (ERM) were proposed by the National Oceanic and Atmospheric Administration [27]. The results showed that the concentration of As was higher than ERL, and that of other HMs was lower than ERL and ERM.

> The concentrations of HMs in the surface sediments of all stations exceeded the background values of coastal sediments [26]. Hg, As, Cu, Pb, Cd, and Zn were approximately 2.7-, 1.9-, 1.6-, 1.5-, 3-, and 1.5-times higher than the coastal background values, respectively, indicating that the surface sediments of the dumping area have been slightly polluted over time. Although the concentrations of HMs in the sediments of the dumping area were higher than the background value, all the HMs except Hg fulfilled the class I standard of the Marine Sediment Quality Standard (MSQS) of China (GB 18668-2002). The quality of sediments in the marine dumping area was required to fulfill the class III standard of the MSQS. Therefore, the HM pollution was within an acceptable range. However, compliance with the sediment quality standards does not mean that HMs have no harm or risk, because the concentrations of HMs cannot fully reflect the pollution state [29], especially, the bioaccumulation and amplification of HMs in marine organisms pose potential health risks to human beings [30,31].

> The coefficient of variation (CV%) was used to explain the changes in HM concentrations in different years. It reflects the average degree of variation of each sampling

station in the samples. If the variation is greater than 0.5, the spatial distribution of HM concentrations in the sediments is uneven, and local point source pollution may exist [32]. The CV% of Hg in the sediments of the Huangmao Island dumping area was greater than 0.5, indicating that Hg entering the dumping area was mainly sourced from external input, that is, dumping. The coefficients of variation of Pb, Cu, As, and Zn were all smaller than 0.5, indicating that these HMs were relatively stable. There is a dynamic process between the release and absorption of HMs in marine sediments. In this process, different physicochemical characteristics (e.g., pH, redox potential, temperature, particle size, and salinity), ocean current, and dumping intensity will cause changes in the CV% values of HMs in sediments [33].

Compared with other dumping areas, HM concentrations in the Huangmao Island dumping area were generally at an intermediate level (Table 2). For individual HMs, the Hg concentration was close to the levels recorded in the Jinzhou Port Dumping Area [34] and Marine Dumping Area outside Jiaozhou Bay [35]. The As concentration was close to the levels recorded in the Hulu Island Port Dumping Area [34] and Tianjin Dumping Area [34]. The Zn concentration was close to the levels recorded in the Hulu Island Port Dumping Area [34], Fangchenggang Dumping Area [36], and Dumping Site at the Taiwan Shelf [37]. The Cd concentration was close to the levels recorded in the Jinzhou Port Dumping Area [34], Huangye Port Dumping [34], Lanshan Port Temporary Marine Dumping Area [38], and Marine Dumping Area outside Jiaozhou Bay [35]. Last, the Pb concentration was similar to those recorded in the Huangye Port Dumping Area [34], Laizhou Port Dumping Area [34], and Fangchenggang Dumping Area [36].

**Table 2.** The concentrations of HMs in sediments from other dumping areas.


#### *3.2. Temporal and Spatial Distribution Characteristics of HMs*

Changes in the concentrations of HMs in the surface sediments in different years are shown in Figure 2. The highest concentrations of Hg, As, Cu, Pb, Cd, and Zn were observed in 2016, 2012, 2013, 2012, 2016, and 2011, respectively. In 2017, the concentrations of all HMs were the lowest, which might be related to the significant reduction in the dumping amount in 2017. From 2011 to 2016, the concentrations of As, Cu, and Pb did not change significantly, that of Zn first decreased and then increased, and those of Hg and Cd increased. Changes in HM concentrations in sediments were affected by multiple factors, such as the pollution level of dumped dredged materials, dumping volume, and hydrodynamic characteristics of the dumping area.

**Figure 2.** Average concentrations of HMs in surface sediments in different years.

Spatial distribution profiles of the HMs in the sediments from the study area are shown in Figure 3. Concentrations of HMs in surface sediments generally showed a high trend in the north and central areas of the dumping area and a low trend in the south of the dumping area. The concentration of Cd was higher in the dumping area but lower outside the dumping area; it reached its highest value at station S3 (in the center of the dumping area) and the lowest at station S5 (in the south outside the dumping area). The concentration of Pb was higher in the central and south of the dumping area but lower in the northwest; it reached its highest value at station S3 and the lowest at station S2 (at the edge of the dumping area). The concentrations of Hg and As gradually decreased from northwest to southeast, reaching the highest at station S2 and the lowest at station S5. The concentration of Cu gradually decreased from north to south, reaching its highest at station S1 and lowest at station S4. The concentration of Zn gradually decreased from north to south in the dumping area, reaching its highest at station S5 and lowest at station S2.

**Figure 3.** Spatial distribution of HMs in sediments.

#### *3.3. Correlation Analysis between HMs and Other Environmental Factors*

The concentrations of HMs in surface sediments are affected by various conditions, including environmental background, marine physical and chemical properties, biological effects, and human activities [39]. Analyzing the correlation between HMs and environmental factors (e.g., oils, TOC, sulfides) is helpful in determining the possible sources of HMs [40].

In this study, the data followed a normal distribution, and the Pearson coefficient analysis was performed using SPSS 25 software. In Table 3, organic carbon had a significant positive correlation with Cu and Zn (*p* < 0.01), indicating that Cu and Zn were closely related to organic matter, which was a main carrier. Additionally, organic carbon and HMs easily form organic matter complexes. Therefore, organic carbon plays an important role in the distribution of HMs. Among the six HMs, Zn was significantly correlated with Cu (*p*< 0.01) and moderately correlated with As and Pb (*p* < 0.05). Hg and Cd were significantly correlated with As but not with the other three HMs, implying that Hg and Cd might have sources similar to As, which differed from those of other metals. Oils and sulfides were not correlated with the concentrations of the six HMs, indicating that they had no effect on HMs in the dumping area.

**Table 3.** Correlation analysis for HMs, TOC, Oils, and sulfides in sediments.


\* Correlation is significant at the 0.05 level (two-tailed). \*\* Correlation is significant at the 0.01 level (two-tailed).

#### *3.4. Pollution Assessment of HMs in Sediments*

Pollution assessments for HMs in sediments have been conducted widely based on several indices, such as the Nemerow pollution index, *Ei r*, *RI*, and *Igeo* [22,41,42].

The Nemerow index was used to evaluate the pollution of HMs in the sediments in the dumping area over different years (Figure 4). In 2012, 2013, and 2016, the Nemerow index exceeded 0.7 but was less than 1. This finding indicated that the sediments in the dumping area had reached light pollution levels in 2012, 2013, and 2016. Sediments in other years were at a clean level. The pollution level in 2016 was slightly higher, which might be related to larger quantity of dumping materials in this year. According to the pollution distribution, the polluted areas were mainly in the northwest (S1 and S2) and central areas (S3) of the dumping area. Additionally, the change in water depth showed that these stations were areas with large dumping quantities, indicating that dumping might have a negative impact on sediments.

The *Igeo* values of As, Cu, Zn, and Pb ranged from 0 to 1 from 2011 to 2016 (Figure 5), indicating that the pollution levels of these four HMs were light pollution. Notably, in 2017, the *Igeo* values of As, Cu, Zn, and Pb were lower than 0, implying that the pollution levels of these four metals were not polluted in 2017, which may be related to the least dumping amount in this year. In most years, the *Igeo* values of Hg and Cd were between 1.0 and 2.0, indicating that Hg and Cd had reached a mild contamination level. From 2011 to 2016, the pollution levels of Hg and Cd showed an increasing trend and were higher than those of the other four HMs. Even in 2017, when the dumping amount was the lowest, the pollution levels of Hg and Cd were rated as light pollution, and the pollution from the other four HMs was not observed. Therefore, special attention should be paid to Hg and Cd before dumping dredged materials.

**Figure 4.** The Nemerow pollution index of HMs in different years.

**Figure 5.** The *Igeo* values of HMs in different years.

Figure 6 shows the integrated potential ecological risk index (*RI*) and potential ecological risk index (*Ei r*) of an individual HM. The *RI* values of HMs in the sediments of the

dumping area were higher than 150 in 2011, 2013, 2014, and 2017, suggesting that HMs had a middle ecological risk. The *RI* values in 2012 and 2016 were 310 and 388, respectively, indicating that the ecological risk of sediments was relatively high. The *RI* values of the dumping area were greater than 150 every year, implying that the potential ecological risk of the dumping area must be paid more attention.

**Figure 6.** *RI* of HMs in different years (**A**) and *Ei r* of HMs in different stations (**B**).

For individual metals, Hg and Cd posed more severe ecological risk than the other HMs, with *Ei <sup>r</sup>* values higher than 40 in each year from 2011 to 2017, especially in 2016, a high risk level (*E<sup>i</sup> <sup>r</sup>* ≥ 160). The *Ei <sup>r</sup>* values of other HMs were less than 40, indicating that the potential ecological risk of other HMs was low. In general, Hg and Cd were the main factors causing the potential ecological risks of sediments in the study area.

At present, extensive ecological risk assessments have been conducted on HMs in sediments based on *RI* [31,43]. In these studies, contamination levels or risk contributions of HMs varied across study areas. For example, Cd was the largest contributor to the ecological risk of sediments [43]. Similar to the results of this study, Lao et al. [31] observed that the ecological risk of Hg in the Beibu Gulf was relatively high, and Tang et al. [42] found that the ecological risk of Hg and Cd in HMs in Daya Bay was also high. Liu et al. [33] found that the ecological risk of Hg in the sediments of the Maowei River aquaculture area was high. Hg and Cd had stronger toxic effects on organisms [44]. The toxicity coefficients (*Tr i* ) of Hg and Cd were 40 and 30, respectively, much higher than those of the other four HMs, which led to a higher potential risk in sediments of the dumping area. Hg and Cd have strong biological toxicity and high bioaccumulation potential, which seriously threaten marine ecosystems and human health. Therefore, to reduce the ecological risk caused by dumping, the monitoring of Hg and Cd in dredged materials should be strengthened before ocean dumping.

The *E<sup>i</sup> <sup>r</sup>* values of HMs at the different stations are shown in Figure 6. In terms of spatial distribution, the order of the potential ecological risk of each monitoring station was S2 > S3 > S4 > S1 > S5. Areas with high ecological risk were mostly located in the northwest of the dumping area, and the ecological risk in the south of the dumping area are low. In the north of the dumping area, the results of ecological risk are consistent with the concentration distribution and the Nemerow evaluation results. In areas (S1, S2, and S3) with frequent dumping, higher concentrations of HMs in sediments were detected, indicating that dumping activities had a significant impact on the marine ecological environment.

#### *3.5. Characteristics of Benthic Community in the Dumping Area*

Fifty-two species of benthos were identified and characterized by the presence of the following groups: arthropods, chordates, annelids, mollusks, echinoderms, nemertinean, coelenterate, and echiuran. Among these groups, arthropods were the most abundant, accounting for 38.5% of all benthos, and chordate, annelid, and mollusks accounted for 25.0%, 15.4%, and 13.5%, respectively. Benthic communities differ in their substrate requirements, and different species dwell indifferent ecosystems [45,46]. After investigating the benthos in Riga Bay in the eastern Baltic Sea, Pallo et al. [47] found that arthropods were more suitable for coarse sand sediment environments. In this study, the similar result was found. Sediments in the Huangmao Island dumping area were mainly silt and the most abundant benthos in this area was arthropods.

A general belief is that filter-feeding benthos are more likely to enrich HMs [48]. However, Quan et al. [49] found that arthropods can easily affect the biological community structure and biomass amount by adsorbing HMs from sediments.

The abundance of taxa, individual density of benthos, and diversity of the benthic fauna are shown in Figure 7. There was little difference in the abundance of taxa, density, and diversity between the central area (S3) and the northwest area (S1 and S2). The abundance of taxa and density of S4 in the southeast corner of the dumping area was low, and those of S5 located to the south of the dumping area were significantly higher than those at the other stations. The response of benthic communities to cumulative dumping was analyzed based on the abundance of taxa, density, and diversity of each station in the dumping area. The abundance of taxa, density, and diversity of benthos at each station were transformed by logarithmic conversion to create a Bray–Curtis similarity matrix and to conduct clustering and MDS sequencing. The results are presented in Figure 8. The benthic fauna community in the study area was divided into four groups at a similar level of 92.9%. S1 and S3 belong to group 1, in the central area and to the northern part of the dumping area. S2 belongs to group 2, in the northwest corner of the dumping area. Groups 1 and 2 had a high similarity. S4 belongs to group 3, in the southeast corner of the dumping area. S5 belongs to group 4, outside the dumping area. Stations in groups 1 and 2 were the areas with the largest dumping amount and higher ecological risk. Stations in groups 3 and 4 were the areas with less dumping amount than group 1 and 2, and pollution and ecological risk were also lower. The temporal and spatial distribution characteristics of benthic communities were highly related to the pollution levels in different dumping areas.

**Figure 7.** The abundance of taxa, individual density of benthos, and diversity for the benthic fauna (Shannon–Wiener index).

#### *3.6. Influence of Dumping Dredged Materials and the Pollution Status*

To minimize the impact of dumping on the marine ecological environment, the dumping distance, hydrodynamic conditions, water depth, and other factors were comprehensively considered during the selection of the location of the Huangmao Island dumping area. In the case of large amounts of dumping and complex types of dredged materials, benthos was affected even during short-term dumping [3]. HMs in dredged materials enter sediments in different ways, causing changes in the sediment environment and affecting the structure and composition of organisms. Jia et al. [50] found that the concentrations of

Cu, Pb, and Cd in sediments had a significant impact on the richness and equity of benthos. Li et al. [51] found that the correlation coefficient between the HM content in sediments and the community structure was the highest, and the content of HMs in sediments was the main environmental factor affecting the community structure of benthos in this area. The HMs in seawater can enter the sediment in different ways, which can change the sediment environment and affect the structure and composition of organisms.

**Figure 8.** Cluster and n-MDS plots based on abundance of taxa, density, and diversity in each station.

The abundance of taxa, density, and diversity have been widely used in the impact assessment of benthos [52]. In this study, there were significant differences in the abundance of taxa and density between the areas inside and outside the dumping area. The abundance of taxa and density outside the dumping area were significantly higher than those inside the dumping area, indicating that dumping dredged materials might affect the benthos. The n-MDS analysis confirmed the observed patterns, separating the S5 stations from the others (S1–S4). Similar to this study, Fonseca et al. [48] also found that the area with more concentrated dumping had more severe pollution and lower richness of benthos. The Nemerow index showed that the central and northern parts of the Huangmao island dumping area (S1, S2, and S3) were polluted, and the *RI* values of HMs in this area were relatively high. Zhang et al. [38] investigated the sediments in the Lanshan Port temporary marine dumping area and found that the concentrations of HMs in the dumping area were significantly higher than those outside the dumping area, corroborating the results of this study.

The results of this study clearly indicate that the source of HMs in the dumping area are mainly from the dumped dredged materials, and dumping had affected the abundance of taxa and the density of benthos, especially in areas with a large dumping quantity and high dumping frequency. Since metal concentrations and evaluation of metal accumulation in benthos were not conducted in this study, the differences in the characteristics of the benthic community between the areas inside and outside the dumping area cannot be directly associated to HM content only. Community structure and composition can also be directly affected by the burial of dumping materials [53]. Other factors, such as dumping amount, dumping frequency, the type of dredged material, and toxicity of the dredged material also change the structure and composition of the benthic community.

#### **4. Conclusions**

The average concentrations of Hg, As, Cu, Pb, Cd, and Zn in surface sediments of the dumping area were 0.080, 15.1, 24.5, 30.8, 0.21, and 97.6 mg/kg, respectively. HMs in the surface sediments of the study area generally showed a trend of high in the northwest and central areas and low in the southeast of the dumping area. All HMs, except for Hg, met the class I standard for MSQS. The Nemerow index showed that sediments in the dumping area were at a light pollution level, and the polluted areas were mainly in the northwest and central areas with large dumping amounts. The values of *Igeo* showed that the dumping area was at a mild contamination level, and Hg and Cd were the major pollutants. The *RI* of the dumping area was relatively high. Hg and Cd are the main factors that cause the potential ecological risk of sediments. Areas with high potential ecological risk and high pollution levels were mainly those that undertake frequent and large dumping amounts of dredged materials.

Benthos identified in this study include arthropods, chordates, annelids, mollusks, echinoderms, nemertineans, coelenterates, and echiurans. Among these groups, arthropods were the most abundant, accounting for 38.5% of all benthos, and chordate, annelid, and mollusks accounted for 25.0%, 15.4%, and 13.5%, respectively. The abundance of taxa and density of benthic organisms showed little difference among the stations within the dumping area but were significantly lower than those of the stations outside the dumping area. It can be inferred that dumping dredged materials in the Huangmao Island marine dumping area had a negative impact on the benthos.

The assessment of HMs in the dumping area indicated that the pollutants in the dredged materials may have harmful effects on benthos. As, Hg, and Cd must be treated before dumping dredged materials. To study the impact of HMs more objectively in dumping dredged materials on marine organisms, metal concentration evaluation in benthic organisms and the toxicity test in sediments of dumping areas should be evaluated in further research. Additionally, the dumping area should be managed by zones to avoid intensive dumping in the northern part.

**Author Contributions:** Conceptualization, W.T. and Z.J.; methodology, W.T. and Z.J.; validation, X.P. and Z.Y.; investigation, Z.J.; data curation, H.Y. and J.Y.; writing—original draft preparation, W.T. and J.Y.; writing—review and editing, W.C. and W.T.; supervision, Z.J. and J.Y.; Funding acquisition, W.T. and J.Y. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by State Oceanic Administration, grant number DOMEP-01-03.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data employed in this study will be available on request to the corresponding authors.

**Acknowledgments:** The authors would like to thank the technicians from South China Sea Environment Monitoring Center, China for sample collection and analysis.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


### *Article* **The Impact of Coastal Geodynamic Processes on the Distribution of Trace Metal Content in Sandy Beach Sediments, South-Eastern Baltic Sea Coast (Lithuania)**

**Dovile Karlonien ˙ e˙ 1,\*, Donatas Pupienis 1,2, Darius Jarmalaviˇcius 2, Aira Dubikaltiniene˙ <sup>1</sup> and Gintautas Žilinskas <sup>2</sup>**


#### **Featured Application: Geochemical analysis can provide valuable information about the local and regional patterns of sediment transport, distribution, provenance, and coasts' conditions.**

**Abstract:** Sandy coasts are one of the most dynamic spheres; continuously changing due to natural processes (severe weather and rising water levels) and human activities (coastal protection or port construction). Coastal geodynamic processes lead to beach sediment erosion or accumulation. The coast's dynamic tendencies determine the changes in the volume of beach sediments; grain size; mineralogical; and geochemical composition of sediments. In addition to lithological and mineralogical analysis of sediments, geochemical analysis can provide valuable information about the local and regional patterns of sediment transport, distribution, provenance, and coasts' conditions. The study aims to assess trace metals' temporal and spatial distribution determined in the sandy beach sediments along the south-eastern Baltic Sea coast (Lithuania) during 2011–2018. The Lithuanian seacoast is divided into two parts: mainland and spit coast. Our results revealed that the dominant group of elements on the mainland includes Ca–Mg–Mn–Ti and on the Curonian Spit Fe–Pb–As–Co– Cr–Ni–Al, which remain unchanged during the years. The analysis included additional parameters such as beach volume, grain size and sorting, and heavy mineral concentration on the beach. The spatial analysis of trace elements indicated that the trace metal content depends on the coastal processes, but it differs in the mainland and spit sea coast. We identified a higher concentration of trace metals in the erosion-dominated areas in all analysed years on the mainland coast. On the spit coast, the trace metal concentration increased in areas associated with relict coarse sand and where the loading of sediments was active on the beach due to the northward along-shore transport.

**Keywords:** trace metals; beach sediments; coastal processes; lithology

#### **1. Introduction**

Trace metals enter the coastal system from the entire Baltic Sea catchment area, which is four times larger than the sea area. The main sources of trace metals besides natural ones in this area are the combustion of fossil fuels (transport and energy production), municipal and industrial sewage management, agriculture, manufacturing processes (pulp and paper, metallurgy, etc.), and military activities (chemical ammunition buried after World War II) [1]. These metals are transported to the sea and coastal areas by rivers, deposited from the air along with precipitation and other pathways.

Sandy beaches are considered important recreational sites and less recognised as highly threatened and fragile natural ecosystems, which may work as natural barriers to pollutants transported by sea [2,3]. Approximately, 50% of coastlines globally are composed

**Citation:** Karloniene, D.; Pupienis, ˙ D.; Jarmalaviˇcius, D.; Dubikaltiniene,˙ A.; Žilinskas, G. The Impact of Coastal Geodynamic Processes on the Distribution of Trace Metal Content in Sandy Beach Sediments, South-Eastern Baltic Sea Coast (Lithuania). *Appl. Sci.* **2021**, *11*, 1106. https://doi.org/10.3390/app11031106

Academic Editor: Mauro Marini Received: 29 December 2020 Accepted: 20 January 2021 Published: 25 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

of sand and gravel, mostly due to river and land sediment transport and deposition [4]. However, the ultimate magnitude of the accumulation/erosion of beach sediments depends also on anthropogenic activities [5]. The textural, lithological, and mineralogical composition of beach sediments depend on many factors: the physical and geographical conditions of the location [6–8], geological framework [9–12], tectonic settings [13–17], provenance [18–22], climate and sea hydrodynamics (waves, tides, and currents) conditions [13,23–27], and the anthropogenic activities [27–31]. Sandy beaches are presently threatened by several forms of environmental degradation, although beach management has traditionally concentrated on geomorphic hazards [32] and recreational use of coast [33], while trace metals are rarely considered [34]. The geochemical analysis of beach sediments together with lithological and mineralogical characterisation can provide valuable information about the local and regional patterns of sediment transport, distribution, and provenance, as well as the conditions of the coast [35].

Global studies related to trace elements analysis on beaches focus on assessing the changes in trace element concentrations in the context of the influence of local pollution sources such as mining sites, urban areas, and industrial complexes [28,29,36–38] or tourism [39–41]. Several studies analysing the migration and distribution of trace elements in coastal environments have been conducted; e.g., the distribution of trace elements concentrations across (from the beach to dunes) [42] and along the coast [38,43–48] and vertical migration [49]. All overviewed studies intended to identify the source or how the particular pollution source affects beach sediments, but the coastal processes were not deeply analysed. The concentration of elements can be affected, as it is known in heavy minerals, by changes in the hydrometeorological conditions [23] or by the season [35,49]. The dominant coastal processes could affect the accumulation of trace elements, which could be of natural origin.

In Lithuania, the formation and dynamics of recent beach sediments have been analysed using lithological, mineralogical, morphometric, and other methods [12,27,50,51]. The nearshore mineralogical and geochemical composition of the Baltic Sea's Lithuanian territorial waters and distribution has already been well analysed [21,52,53]. However, the detailed geochemical composition of beach sediments in relation to active coastal processes has not been previously investigated, except for some local studies [54,55]. This study aims to assess the temporal change of the trace metal content in sandy beach sediments and the dependence on coastal lithomorphodynamical processes.

#### **2. Study Area**

The Lithuanian coast is divided by the 1.1 km wide Klaipeda Strait, where Klaip ˙ eda ˙ Port is located into two parts: the Curonian Spit (hereinafter—the spit) sea coast (51 km) and the mainland coast (39 km) (Figure 1). Beyond 1991, when the Curonian Spit National and Seaside Regional Parks were established, the major part of the Baltic Sea coast in Lithuania (about 70 km long) has acquired the status of a protected area. Several coastal sectors with different characters can be distinguished: a technogenic coast that predominates near to Klaipeda and Šventoji Ports, i.e., areas of waste water disposal from the B ˙ uting ¯ e oil ˙ terminal, and the Mažeikiai oil processing plant, and protected areas such as the Curonian Spit National Park, the Seaside Regional Park, the Baltic Sea Talasological Reserve, and the Buting ¯ e Geomorphological Reserve (Figure 1). ˙

**Figure 1.** Study area. (**1**) surface sand sampling site, (**2**) main settlements, (**3**) D6 oil platform (Ru), (**4**) Buting ¯ e oil terminal, ( ˙ **5**) conservation areas, (**6**) borders, (**7**) offshore and nearshore dumping area, (**8**) beach nourishment site in Palanga, and (**9**) storage site of dredged bottom sediments.

The Holocene clastic deposits occur along the entire length of the Lithuanian coastal zone. The Lithuanian coastal sediments are composed of quartz, K–feldspar (orthoclase, microcline), plagioclase (albite, anorthite), carbonates (dolomite, calcite), mica (biotite, muscovite), and clay minerals (illite, chlorite, kaolinite, montmorillonite, glauconite, and vermiculite) with admixture of heavy minerals [21,50].

On the mainland coast, the average 20–100 m wide beaches are backed by foredunes 4–12 m in height or by moraine cliffs 5–24 m in height. The southern part of the mainland coast structure between the 25th and 31st km (distance from Latvia-Lithuania border) is characterised by glacial (moraine) deposits formed during the Late Pleistocene and that, in most cases, occur on the abraded cliff coast. The first moraine cliff is located between the 25th and 27th km on the mainland coast near Šaipiai (Figure 1). The cliff is 2.1 km long and 5–8 m in height. The second moraine cliff (Olando kepure) is 950 m long and 24 m ˙ in height at its highest point [56]. The highest cliffs are between the 30th and 31st km at Karkle. On the Curonian spit sea coast, the average width of the beaches varies from 32 to ˙ 75 m, and the relative height of the foredune changes from 5 to 16 m [12].

The Baltic Sea is nontidal (amplitudes reach 3.5–4.0 cm); therefore, a wind-wave regime dominates. The cold (autumn–winter) season is marked by most days with strong winds. The annual mean wind speed is 4.7 m s−<sup>1</sup> and the wave height is 0.65 m. During the strongest storms, the wave height varies from 5 to 6 m [56]. The prevailing westerly (SW, W, NW) winds and waves are dominant in the coastal zone and generate the alongshore sediment transport from the Sambian Peninsula to the end of the Curonian Spit and along the mainland coast [24,27,57]. Currently, the alongshore sediment transport is disrupted by the Klaipeda port gate [58]. ˙

#### **3. Materials and Methods**

In the analysis, we focus on trace elements which are considered of anthropogenic origin. The macroelements, lithological and morphological parameters are used for the data interpretation and definition of the causes.

Sediment samples were collected along the entire Baltic Sea coast of Lithuania in 2011, 2014, and 2018 (Figure 1). During the collection of samples, the sea and wind conditions were relatively constant. In the coastal zone, the most active processes occurred in the surf and swash zone; for that reason, it was decided to take samples from the middle of the beach [26]. Totally, 43 composite surface sand (0–5 cm) (from 5 subpoints) samples were collected at equal distances of 3 and 5 km and in plastic containers delivered for laboratory analysis [27]. In the laboratory, sand samples were air-dried and split to 100 g of sediment. The samples were mechanically sieved 15 min on a vibratory sieve shaker Fritsch Analysette 3 Spartan Pulverisette 0 using a set of 11 sieves. After determination of size fraction, statistical grain-size Falk and Ward method mean (d, mm), sorting (*So*) parameters were calculated using the GRADISTAT 8.0 software [59].

To characterize the relative concentrations of heavy minerals, a Bartington MS3 field scanning sensor was used for rapid and effective measurements of low-field volume magnetic susceptibility [27]. Sandgren and Snowball [60] point out that bulk magnetic susceptibility (MS) is a good indicator of allochthonous mineral matter in sediments. Measuring MS helps to determine the net contribution of ferromagnetic and paramagnetic minerals in sediments. Heavy mineral-rich (ρ > 2.90 g/cm3) sediments have ferromagnetic and paramagnetic properties, and high magnetic susceptibility values, depending on the predominant iron content. Quartz-rich minerals (ρ < 2.65 g/cm3) have diamagnetic properties. Quartz-rich sands (dominate quartz, feldspar, carbonate, and mica group minerals) have weaker positive magnetic susceptibilities values of κ < 3.0 μSI, heavy mineral-rich sand (elements like Ti, Cr, Mn, Fe, Co, Ni, and Cu can sometimes result in magnetism) κ values range from 30 to 150 μSI and higher values κ > 150 μSI are typical for heavy minerals with Fe, Ni, and Co elements [26,61,62]. MS measurement method is helpful to detect ferromagnetic minerals when their concentration in sediments is deficient [60]. Magnetic susceptibility and grain sizes of beach sediments belongs of the provenance, geologic framework, alongshore sediment transport, deposition and coastal processes (erosion/accretion), etc. The grain size composition also might have contributed to the difference in proportions of ferromagnetic, paramagnetic, and diamagnetic minerals [63]. Magnetic minerals are known as important sources of trace elements in sediments.

The beach sediment volume (*Q*, m3/m) was calculated for each profile based on repeated cross-shore levelling once per year. The changes in sediment volume comprise the changes in volume of the coastal profile from the foredune lee side, where the vertical variability is negligible during observation, to the intersection with the mean sea level [12]. Total beach sediment volume was counted with the formula: *Q = (Qi + Qi +* 1*) Li*/2, where *Q*—sediment volume (m3) at the coastal segment; *i* = 1, 2, 3, ... number of transects; *Qi*—sediment volume at the separate coastal cross-section profile (m3/m); and *Li*—distance between levelling cross-section profile lines [64].

For the geochemical analysis, the dried samples were ground using an agate mortar and pestle. Prior to each sample's homogenisation, the agate mortar and pestle were washed with deionised water and dried twice. Geochemical analyses of the samples were performed at the Bureau Veritas Commodities Canada Ltd., laboratory. 0.5–2.0 g of bulk sample digested after application of modified aqua regia (1:1 HNO3: HCl) solution for low to ultralow determination of soil and analysed with an inductively coupled plasma mass/emission spectrometer (ICP–MS/ES). In this study, we mostly analysed trace elements that might originate from the anthropogenic activities common in coastal systems as fossil fuel burning, sewage discharge, metals common in ship ports, etc. and monitored at the national monitoring program [65]. We also selected macroelements that could help to describe the origin of sediments. The results are given in ppm for elements As, Cu, Cr, Co, Mn, Ni, Pb, and Zn; in ppb for Hg; and in percentage for Ca, Mg, Fe, Al, and Ti. Analytical quality was monitored in each batch of samples by repeated analyses, recovery of spiked samples, and analysis of a certified reference material (OREAS45EA and DS11), and duplicates and blanks were used to assure the quality of the analysis. The reaction mixture was chosen to evaluate the labile trace elements dissolving sulphide/oxide type minerals to exclude the elements incorporated in the silicate lattice as metals from anthropogenic sources tend to be more mobile than those from pedogenic or lithogenic sources [43,66].

Descriptive and multivariate statistical methods were applied to analyse the results. Correlation analysis using Pearson's coefficient (linear relation, significant *p* > 0.01 or 0.05) and principal component analysis (PCA) were applied to transform the correlation matrix to identify the relationship between the metals [31]. For the PCA analysis, we used the Varimax rotation method to derive more reliable information on the distribution of the weights of the variables on a factor, and loadings higher than 0.5 were considered. Statistical programs IBM SPSS Statistics 22.0 and PAST 3.24 were used.

The concentration of trace metal in each site was compared with the median estimated of all samples at that year *Kk* = (*Kn*)/(*Mn*), where *Kn*—concentration of the element *n* and *Mn*—the median concentration of the element *n*. Median presents the central or typical value in a set of data and is weakly dependent on minimal and maximal values and outliers, this approach has been used and developed in a few studies [48,52,67]. This analysis helps to compare the loading of analysed elements among sites; the lower limit of the anomaly is considered as 1.5 [52]. Following a multi-element index, the integrating and averaging data were estimated, *Kd* = Σ*Kk*/*n*, where *Kk*—is the concentration ratio for a specific element and *n*—number of elements. We assumed that *Kd* values higher than 2 indicate an anomaly concentration in the site [68].

Other studies use concentration ratios to determine trace element accumulation in sediments and focus on anthropogenic pollution [43,53,69–71], the determination of background concentrations in such studies is essential [71]. The main concept to identify the anthropogenic impact is to compare the concentrations of elements measured in uncontaminated sites—to establish a local baseline or, as in most studies, compare sediment element concentrations with preindustrial levels such as average shale [72] or average crustal value [73]. Using the average shale or earth's crust concentrations, the local geochemical framework is being ignored, which might lead to misinterpretation of anomaly concentrations in the analysed region. Second, the shale concentration mostly represents fine grain sediments and earth's crust—coarse sediments, meanwhile the background concentration represents sites in similar mineralogical and textural environments [71]. In this study, we analysed beach sand, dynamic environment and usage of shale or earth's crust as the background concentration might miss an important pattern in the distribution of trace elements. For that reason, we chose to use the values calculated during our study. We also considered the values as a background—calculated from 76 samples of aeolian sediments collected in the western region of Lithuania located close to our studied area [74]. However, this area is remote from the sea—where the sediments are affected by both wind and water; thus, we considered not to apply these data in the current study.

#### **4. Results**

#### *4.1. The Lithology of Beaches*

The sandy beaches on the Lithuanian sea coast are composed of fine-medium grain sediments (*d* = 0.27 mm, *σ* = 0.07 in 2011; *d* = 0.29 mm, *σ* = 0.09 in 2014; *d* = 0.29 mm, *σ* = 0.11 in 2018). Although in 2011, fine and medium sand predominated on the beaches, in 2014 and 2018, coarse sand was already detected in several places (Figure 2). In 2011, fine sand dominated in 23 and medium sand in 20 sites. In 2014, fine sand was determined in 20 sites, the medium—dominated in 21 places, and coarse sand—in two sites—73rd and 131st. In 2018, fine sand was indicated as well in 21 places, medium sand in 19 places, and coarse sand in 3 places—61st, 73rd, and 133rd. The appearance of coarse sand at Olando kepure, north of Klaip ˙ eda port pier, is related to the intensified local erosion processes. ˙ Erosive processes are local and usually occur in specific stretches of the coast, e.g., where moraine cliffs predominate, or are affected by hydrotechnical construction.

**Figure 2.** The distribution of grain-size (d, mm) and sorting (*So)* of the beach sediments along the coast of the south-eastern Baltic Sea coast (Lithuania) in 2011, 2014, and 2018.

In all investigated years, the surface sand on the mid-beach was very well sorted (*So* < 0.35) and well sorted (0.35 < *So* > 0.50) (Figure 2). Moderately well-sorted (0.50 < *So* > 0.70) or moderately sorted (*So* > 0.70) sand was in areas with a medium or coarse sediment fraction. In 2011, moderately sorted sand was identified in 9; in 2014—in 13, and in 2018—in 14 sites. The distribution of beach sand particles shows that the average particle diameter decreases from south to north on both the mainland and the Curonian Spit coast (Figure 2).

#### *4.2. Magnetic Susceptibility of Beach Sediments*

The bulk MS values of the sediments on the investigated coast ranged between 13.9 and 357.7 μSI (mean *κ* = 64.4, *σ* = 58.7) in 2011, between 16.5 and 804.8 μSI (mean *κ* = 139.1, *σ* = 161.6) in 2014, and between 7.9 and 271.8 μSI (mean *κ* = 77.7, *σ* = 60.2) in 2018 (Figure 3). The MS values differed between the mainland and Curonian Spit coast, i.e., in 2011 and 2014, the MS values measured higher on the mainland coast, and in 2018, the difference in the values faded. The anomalous MS values were determined at the 7th, 21st, 127th, and 181st.

**Figure 3.** The distribution of magnetic susceptibility values (κ, μSI) of the beach sediments along the coast of the southeastern Baltic Sea coast (Lithuania) in 2011, 2014, and 2018 (the dark green colour field indicate heavy minerals, green—heavy mineral-rich sand and light green—quartz-rich sands).

The MS values differed between the mainland and Curonian Spit coast. In 2011 and 2014, the MS values were higher on the mainland coast, and in 2018, the difference in the values faded. The higher values of magnetic susceptibility were detected in the beach sediments, enriched with perimagnetic and ferromagnetic minerals. The anomalous MS values were determined at the 7th, 21st, 127th, and 181st. Magnetic susceptibility reflects different coastal processes and magnetic mineral sources. Magnetic susceptibility of beach sediments increases greatly close to the provenance and decreases moving away from source.

#### *4.3. Beach Sediment Volume*

The beach sand volume (*Q*, m3/m) compared between 2011 and 2014 increased by in almost all sites on the mainland coast except in the 25th, 41st, and 49th (Figure 4). However, on the Curonian Spit coast, the sediment content on the beaches in 2014 compared to 2011 decreased in the end of the spit (79th and 91st–109th) and in Juodkrante–Pervalka section ˙ (115th, 127th, 133rd, 139th), higher in the 141st, 161st, and 175th sites, the highest increase in sediments was in the south of Nida (157th, 169th, 171st, and 181st).

**Figure 4.** The changes in beach sediment volume (*Q*, m3/m) distribution along the coast of the south-eastern Baltic Sea coast (Lithuania) in 2011, 2014, and 2018.

Comparing 2018 and 2014 sediment volume on the beaches, on the mainland coast the highest increase was determined in the 31st, 49th, and 61st, and the significant decrease was determined in the 13th, 37th, and 51st sites, but mostly in the 73rd. On the Curonian Spit coast, the sediment volume increased mostly at the end of the Curonian Spit (79th–97th, 109th, 115th) and 141st. The considerable decrease in sediments on the beaches of the spit was estimated on Juodkrante–Pervalka (121st, 127th, 131st), 169th, and 181st sites. ˙

#### *4.4. Descriptive Statistics of Trace Elements*

The descriptive statistics for the 3 years of the analysed trace elements (Cu, Pb, Zn, Ni, Co, Mn, As, Cr, Hg, and Cd) and macroelements (Fe, Al, Ca, Mg, and Ti) are provided in Table 1.

The comparison of the average concentrations of the trace elements in the analysed years revealed that the higher average concentrations of most elements were measured in 2014. In 2011, the higher concentrations of As and Cr compared to other years were determined, while in 2018, only for Mn. Additionally, in 2014, Hg was measured in sediments collected on 18 sites; in 2011, Hg was detected only on eight sites; and in 2018, Hg was found on six sites.

#### *4.5. Distribution of Trace Elements Along Coast*

We assessed the distribution of the average concentration of trace elements and the medium value ratio (*Kk*) along the coast. The pattern of the ratio differed among elements. The *Kk* of As were estimated only on the Curonian Spit coast where the ratio exceeded 1.5 in six sites on the distal end of the spit (79th, 85th, and 103rd–111st) and Pervalka–Preila section (145th) in 2011, in four sites (109th, 111th, 131st and 145th) in 2014, and in only one Juodkrante–Pervalka section (133rd) in 2018. We estimated higher than 1.5 Kk values ˙ for Cr in 2011 in seven places, a significant majority at the distal end of the Curonian Spit (79th–111th); in 2014, already in 12 sites (79th, 91st, 111th, 131st, 145th, 157th, 171st, and 175th), and only in one site (51st) on the mainland coast (Figure 4). In 2018, as in previous years, we indicated the highest Cr concentration at the distal end of Curonian Spit (79th–109th) and in the coast stretch southern from Juodkrante (133rd, 139th, 145th, ˙ 151st, and 181st) (Figure 5).


**Table 1.** Trace element concentrations of the surface sediments collected on the beach along the coast of the south-eastern Baltic Sea coast (Lithuania) in 2011–2018 (values in ppm unless otherwise indicated).

**Figure 5.** *Cont*.

**Figure 5.** *Cont*.

**Figure 5.** *Cont*.

**Figure 5.** The distribution of concentration ratio (*Kk*) of (**a**) Cu and Mn, (**b**) Cr and Zn, (**c**) As and Co, and (**d**) Ni and Pb along the coast of the south-eastern Baltic Sea coast (Lithuania).

The Cu concentration ratio exceeding or equal to 1.5 was estimated in five sites (1st, 7th, 49th, 55th, and 73rd) on the mainland coast in 2011; in 2014, in three sites (51st, 127th and 157th), and in 2018, in four sites (49th, 61st, 71st and 133rd). The *Kk* of Cu in 2014 exceeded four in five sites located near Šventoji (1st, 7th, 11th, 13th, and 25th) in the north part of the mainland coast and two sites on the Curonian Spit between Juodkrante and ˙ Pervalka (133rd and 141st sites) in 2018. The ratio *Kk* of Pb did not exceed 2.5 compared with Cu. In 2011 and 2014, these ratios were higher than 1.5 only in the 145th site in Pervalka and in none of the sites in 2018.

The higher concentration ratios of Mn were estimated on the mainland coast opposite to Cr. In all analysed years, higher Mn concentration ratios were estimated in the cliff area (49th–55th). Higher Mn ratio in 2014 and 2018 was also estimated in the site at 7th and also in the 13th and 73rd sites in 2018.

Zinc showed no clear pattern or difference between the mainland and Curonian Spit coasts (Figure 5). Overall, the highest Zn concentration ratios were determined on the mainland coast at the cliff area – in the 67th site in 2014 and the 49th in 2018. In 2011, the Zn concentration ratio was higher than 1.5 in eight sites (7th, 79th, 85th, 103rd–111th, and 145th), in 2014 in four sites (1st, 49th, 67th, and 71st) and in 2018, in six sites (7th, 49th, 61st, 71st, 85th, and 133rd).

Cobalt pattern differed from other elements because its concentration did not variate so much among the years. The concentration ratio exceeded 1.5 in 2011 only in two sites (111th and 145th); in 2014, in the sites 43rd, the 131st, and 145th; and in 2018, only in the 133rd site. The Ni as well as Co, Pb did not show a clear pattern, its ratio exceeded 1.5 in the 111th and 145th sites in 2011; 131st and 145th in 2014; and 55th, 61st, and 133rd sites in 2018.

To understand the overall distribution pattern of the analysed elements in 3 different years, we estimated and compared the mean of the concentration ratios (*Kd*) in 2011, 2014, and 2018 (Figure 6). In 2011, slightly higher mean concentration ratios were indicated in the area where coastal erosion is active; 1 km northwards from Šventoji port (7th site), in the cliff areas (49th and 55th) and near the Klaipeda strait (73rd site). On the Curonian Spit ˙ coast, the distribution pattern of the mean concertation ratio is more pronounced than on the mainland coast, and increases northwards (highest peak at the 111th site).

In 2014, the trend of *Kd* on the mainland coast distinguished from the pattern in 2011 and 2018. The mean concentration ratio was higher than 2 in 11 sites from 18 on the mainland coast and increased northwards. The mean concentration ratio exceeding 2 was determined in three sites located at the vicinity and northwards from the Šventoji port. On the Curonian Spit, the trend was opposite to the other years—the mean ratio slightly decreasing northwards.

In 2018, on the mainland coast, the estimated *Kd* ratio values tend to decrease northwards, the highest values were determined in the cliff area (49th and 61st sites), and only in three sites was higher compared to previous years (49th, 61st, and 71st sites). On the Curonian Spit coast, the trend of the mean ratio increases northwards; however, it is mostly expressed from the 121st to 79th sites and from the 181st to 127th site, and the highest rations were determined in the section between Juodkrante and Pervalka (127th to 133rd ˙ sites and 141st site). In general, the trend is similar to 2011.

**Figure 6.** The distribution of the mean concentration ratio (*Kd*) of all analysed trace elements along the south-eastern Baltic Sea coast (Lithuania).

#### *4.6. Correlation Analysis*

The correlation analysis of trace elements and lithological and geomorphological factors showed that Cu content tends to accumulate in beach sediments where erosion processes are active as in 2011. In addition, it is depended on the sorting of the sediments (*r* = 0.40) (Table 2) and was lower in the well-sorted sediments and negatively correlated with beach volume. The copper concentration positively correlated with the MS values in 2011 and 2014 (*r* = 0.59 and *r* = 0.46).

**Table 2.** Pearson's correlation coefficient between trace metals and mean grain size (d, mm), sorting coefficient (*So*), beach sediment volume (*Q*, m3/m), and magnetic susceptibility (MS, μSI).


\*\* Correlation is significant at the 0.01 level (2-tailed). \* Correlation is significant at the 0.05 level (2-tailed).

The manganese content in all analysed years negatively correlated with beach sediment volume (*r* = −0.46, r = −0.45, and *r* = −0.45) and positively correlated with MS values (*r* = 0.79, *r* = 0.83, and *r* = 0.36). In 2011, the Mn concentration positively correlated (*r* = 0.38) with the sorting coefficient and in 2018 with grain size (*r* = 0.39). The zinc concentration positively correlated with MS values in 2014 and 2018 (*r* = 0.45 and *r* = 0.33). In 2011, the concentration positively correlated with beach sand volume (*r* = 0.49). In 2018, the As content positively correlated with grain size and sorting coefficient (*r* = 0.45 and *r* = 0.34).

The grain size positively correlated with sorting coefficient in all investigated years; in 2018, grain size also positively correlated with MS values and negatively with beach volume (Table 2). In 2011, the sorting coefficient of beach sediments positively correlated with MS values. The measured beach volume negatively correlated with MS values in all years, but only in 2014 and 2018 it was statistically significant.

#### *4.7. Multivariate Analysis*

The PCA analysis was used to reduce the number of analysed variables and to identify the different factors that controlled the distribution of the elements in the beach sediments. For further analysis, we added macroelements as, Al, Ca, Fe, Mg, and Ti, which helped to explain the origin of the trace elements. The results present principal components whose eigenvalues were above 1.

The PCA analysis of 3 years data revealed that there are two dominant groups of elements that are defined by the eigenvalues and the correlation coefficients (Figure 7). In 2011, three components were extracted, the first group (PC1) Fe–Co–Al–Cr–Ni–Zn–As– Pb (explains 52.2% of the total variance with the highest eigenvalue 6.8), second (PC2) Mn–Ca–Mg–Ti (eigenvalue 3.4; explains 26.2% of the total variance), and third (PC3) Cu (eigenvalue 1.27; explains 9.7%). In 2014, only two components were extracted, the first group consisting of the following elements Fe–Co–Ni–Cr–As–Al–Pb (eigenvalue 5.3; explains 40.9% of the total variance) and second Mn–Ca–Mg–Ti (eigenvalue 4.2; explains 32.3% of the total variance). In 2018, the element distribution between groups was similar as in other years—PC1: Fe–Pb–As–Co–Cr–Ni–Al (eigenvalue 5.2; explains 40.3% of the

total variance), PC2: Mn–Ca–Mg–Ti–Ni–Al (eigenvalue 4.7; explains 36.3% of the total variance), and PC3: Cu (eigenvalue 1.12; CV 8.6% of the total variance).

**Figure 7.** The loadings of analysed elements in different component groups after principal component analysis (Varimax rotation) in beach sediments of south-eastern Baltic Sea coast (Lithuania) in 2011, 2014, and 2018.

The PCA results showed that the elements from the first component group in 2011, 2014, and 2018 dominated in the Curonian Spit. The second group described sediments from the mainland coast (Figure 8).

**Figure 8.** Principal component analysis after normalized varimax rotation, and factorial scores showed according to geographical location (Curonian Spit and mainland coast) in 2011, 2014, and 2018.

#### **5. Discussion**

It was found out that the trace element concentrations of beach sediments depend on the coastal processes and nonsignificantly varied among years, except for Cu and Zn in 2014. The Mn and Cr showed a clear distribution pattern along the coast during the years. Manganese concentration ratio was higher on the mainland coast, and it tended slightly to decrease northwards. The Cr concentration ratio was higher in Curonian Spit and increased at the distal end of it where the unloading of the sediments is active. These tendencies indicate that the alongshore sediment transport play a significant role in the distribution of the elements along the coast. The decrease in grain size in the north direction shows that the accumulation processes are conditioned by alongshore sediment transport directed from south to north (Figure 2). The measured amounts of beach sediment volume agree with these statements. Poorly sorted sediments were found in anthropogenically affected sections of the coast (Buting ¯ e, Šventoji, Palanga, and Klaip ˙ eda) and in places where the ˙ sand grain size depends on the geological framework (Juodkrante–Pervalka stretch). ˙

The higher Mn concentration was determined in the sites with higher heavy mineral content and with a reduced beach sediment volume. Finer light mineral fractions such as quartz, feldspar, and muscovite mica are usually washed away, and the fine but heavier particles related to heavy minerals co-occur with coarser quartz particles [35,61]. Manganese according to the PCA analysis is associated with the second component group—carbonates (Ca and Mg), and this association remained during the investigation years. The dominance of carbonates can be explained by the deposits of detrital dolomite and biogenic calcite found on the mainland coast [21,36]. The greatest source of calcite and dolomite was found at Klaipeda and gradually decreased towards the south of Palanga [ ˙ 21]. The area southward from Klaipeda strait is less abundant in calcite [21]. ˙

The Mn distribution pattern also showed that its main source is at the moraine cliff area (Šaipiai–Olando kepure). Its concentration increased during the years probably due to ˙ more active erosion on the mainland coast. Carbonate rocks also tend to form compounds with cations of divalent metals such as Cu, Zn, Mn, Sr, Pb, etc. [54]. The Zn pattern shows that its concentration was also the highest close to the moraine cliff area. In addition, enlarged Zn concentration in 2014 and 2018 relates to higher heavy mineral content. The results indicated that its concentration is also linked to the erosion processes. The PCA analysis revealed that Zn concentration in 2011 is associated with sediments from the Curonian Spit, while in 2018 and 2014, there was no concrete association with any of groups. In 2014, the content of Zn was significantly higher than in other investigated years. This suggests that there might be additional sources besides natural ones. The cliff area (49th—61st sites) 30 years ago was an active military zone [75], which also could have affected sediments resulting in higher content of Zn. The trace elements emitted in the past could migrate into deeper layers and in erosive areas could re-enter the environment after storms [76]. The Zn, Mn, and Cu content is higher in the 71st and 73rd sites, which overlap with the dumping sites of the sediments dredged from the Klaipeda Strait (Figure 1). The ˙ dredged sediments are more enriched with trace elements than beach sediments [53].

Higher than average MS values are associated with an increase in heavy minerals, whose content depends on the geological framework, which could enhance after a storm [26]. On the distal end of Curonian Spit the sediment accumulation prevail due to sediments delivered by the alongshore sediment transport [27,57]. The Curonian Spit coast beaches are enriched with glauconite (greensand) originating from Neogene–Paleogene deposits from the Sambian Peninsula [27,48]. The trace metals tend to attach to clay particles such as glauconite, mica, and biotite, which are transported northwards and unload at the distal end of the spit [21,48,50,77–79]. This relation is confirmed by the PCA analysis results, it means Cr together with Pb, Ni, Co, and As in all analysed years form a group of metals associated with Al and Fe (Pb–Ni–Co–Fe–As–Cr–Al) which constitute the composition of glauconite [80]. Therefore, despite the sediments are enriched in trace elements in this part of the spit, they are characterised by decreased MS values as a result of coverage with quartz sand from aeolian processes.

As declared previously, the positive correlation with grain size and MS value indicated the coastal erosion. However, on the Curonian Spit, there is no active erosion [12,81]. The only anomaly of coarse sand in the Curonian Spit is located near Juodkrante—Pervalka ˙ site. The long-term sediment grain-size study recorded the relict origin of the anomaly and general stability of coastal sections of the spit [12]. This is in line with our lithological analysis. The concentration of analysed elements was determined higher in this area especially in 2011 and 2014 at 145th site and in 2018 at 131rd—133th sites. Since these coast stretches of the Curonian Spit were also richer in heavy minerals, this could cause the higher concentration of elements [82]. The correlation analysis revealed relationship related to greater elements' concentration (Ni and Co in 2014 and Ni, Co, Pb, and As in 2018) in coarser beach sands.

In 2014, the analysed trace elements' content was higher than the concentrations determined in 2011 and 2018, specifically, the loading of Cu and Zn on the mainland coast. Similar to Zn, the increased Cu concentration was associated with a higher content of heavy minerals only in 2011 and 2014. However, in 2014, Cu concentration significantly increased to the north of Šventoji Port that could suggest an anthropogenic source. In this case, the sediments' geochemistry could be affected by the Šventoji Port reconstruction in 2011–2012, when the bottom sediments were dredged from the entrance channel and stored on the beach [83], and later washed away after the storm "Xaver" in 2013. The bottom surface sediments in the Šventoji Port basin were enriched with Pb, Ni, Cu, and Zn [84]. Additionally, the sediment in the north of Šventoji has greater absorption capacity due to the peat layer, which is exposed after stormy weather event [85].

The study has some limitations regarding ambiguous evidence of the anthropogenic impact. Our hypotheses are based on literature analysis. For example, our results show anomalies of trace elements southwards from Juodkrante on the Curonian Spit coast. These ˙ anomalies could be also consequences of the local fisheries activities due to possible oil leakage or accidents at a D6 oil platform, coastal protection construction of tires, or cement block at the Sambian Peninsula in the provenance region of the spit sediments [27,67].

#### **6. Conclusions**

Distribution analysis of the trace elements on the south-eastern Baltic Sea coast indicated that the concentration mainly depended on the coastal processes (alongshore sediment transport, coastal erosion, and sediment accumulation). The differences in trace element concentrations and composition between the Curonian Spit and the mainland coast indicated various sources. The trace metal concentrations on the mainland coast mainly increased in the areas with active erosion processes when heavy minerals were exposed. On the Curonian Spit, trace element anomalies are associated with relict sands. The elements such as Cr and as tend to accumulate at the distal end of the spit where sediments most actively accumulate. Since the coast is a dynamic environment, the elements' concentration at the same sites may vary from year to year. For example, in 1 year, severe weather events may lead to more intense coast erosion. In another year, when calm weather is favourable for accretion, these sediments might be covered with fine quartz sand or reworked and redeposited. However, our results showed that beach sediment geochemical composition remains invariable in space and time. The anomalous concentration of Cu and Zn on the coast indicated the possible anthropogenic impact on sediment geochemical composition (such as former military activities and beach nourishment with dredged sediments from the port).

**Author Contributions:** Conceptualization, D.K. and D.P.; methodology, D.K., D.P. and D.J. and G.Ž.; formal analysis, D.J., D.P., and G.Ž.; investigation, D.K., D.P., and A.D.; data curation, D.K., A.D., D.J. and D.P.; writing—original draft preparation, D.K.; writing—review and editing, D.P., D.J., A.D. and G.Ž.; visualization, D.P. and D.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research is a part of a project (duration 2017–2021) funded from the public budget of the Republic of Lithuania.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Acknowledgments:** The authors would like to thank the reviewers for their constructive comments, which helped improve our manuscript. Authors are also very thankful for student Anna Cichon-Pupienis for fine tuning the manuscript.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## **Sources and Metal Pollution of Sediments from a Coastal Area of the Central Western Adriatic Sea (Southern Marche Region, Italy)**

**Federico Spagnoli 1,2,3, Rocco De Marco 1, Enrico Dinelli 4, Emanuela Frapiccini 1,\*, Fabrizio Frontalini <sup>5</sup> and Patrizia Giordano <sup>6</sup>**


**Abstract:** Sediments represent a critical compartment of marine coastal ecosystems due to the toxic and long-lasting effects of the contaminants buried therein. Here, we investigated the properties of surficial sediments in front of the Southern Marche Region coast (Central Adriatic Sea, Italy). The grain size of the surficial sediments was determined by X-ray sedigraphy. TN and OC contents were determined by elemental analysis. The concentrations of Al, Fe, Mg, K, S, Ca, Ti, P, Na, Mn, Mg, Li, As, Ba, Ga, Pb, Sr, and Zn were determined by ICP-OES to evaluate their spatial patterns and temporal trends. A Q-mode Factor Analyses was applied and resulted in the identification of three compositional facies (Padanic, Coastal, and Residual) characterized by common biogeochemical, mineralogical, sedimentological properties, transport pathway, and source. Some pollution indicators, such as the enrichment factor, the geoaccumulation index, and the pollution load index were calculated to assess the deviation from the natural background levels. The results showed a pollution by As and Ba due to the human activities in the 20th century. Furthermore, a general decreasing of Al, Ti, P, Co, Cr, Cu, Ga, Ni, Pb, Sc, V, and Y concentrations from the background levels suggested a change in the sedimentation processes during the last decades.

**Keywords:** pollution indicators; surficial sediments; central western Adriatic Sea

#### **1. Introduction**

The biogeochemical and sedimentological properties of bottom marine sediments are the results of complex interactions between their composition, sourcing areas, transport, depositional, and early diagenesis processes [1]. Additionally, anthropic activities by discharging contaminants can further influence the natural composition of the sediments. Considering all the factors affecting the biogeochemical and the sedimentological composition of the bottom marine sediments, information about the source area, the sedimentological processes, and the anthropic inputs can be inferred [2–7]. Wastewater discharges such as domestic wastes, drugs, fertilizers, and zootechnical byproducts represent the major anthropogenic contributions in marine coastal area [8]. Anthropogenic chemicals can be introduced into the marine environments by different sources (i.e., effluent treatment plants, accidental discharges, dumping, riverine inputs, surface runoff,

**Citation:** Spagnoli, F.; De Marco, R.; Dinelli, E.; Frapiccini, E.; Frontalini, F.; Giordano, P. Sources and Metal Pollution of Sediments from a Coastal Area of the Central Western Adriatic Sea (Southern Marche Region, Italy). *Appl. Sci.* **2021**, *11*, 1118. https:// doi.org/10.3390/app11031118

Academic Editor: Mauro Marini Received: 22 December 2020 Accepted: 22 January 2021 Published: 26 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

atmospheric deposition, etc.) and then accumulate in sediments [9–15]. In coastal marine sediment, however, heavy metals might be sourced by anthropic activity or by natural sources [8,16] and could represent a serious hazard due to their toxicity, bioavailability and persistence [17,18]. Because of the adsorption, hydrolysis, and co-precipitation of metal ions, most of the water column metals are deposited in the marine sediment and only a small portion of these ions remains dissolved in the water column [19]. Then, marine sediments enriched in metals can release them by various remobilization processes to the water column [1].

Marine sediments are preferred over seawater for monitoring environmental quality because pollutants are much higher and less variable in time and space than in seawater, and they reflect in an integrated manner the pollution level in the area [7,20–22]. Marine sediments act not only as a reservoir of contaminants, but also serve as a source of toxicants to marine fauna. This can be due to the ingestion of bottom or resuspended sediment particles [23,24], or to the adsorption of solutes present in the pore waters and on the bottom. The solutes present near the bottom can be the consequence of diffusive or bio-irrigation fluxes or resuspension [1], resuspension that can be due to natural (by storms, waves, tidal currents) or anthropogenic (dredging, trawling) processes [25,26]. For these reasons, sediments are commonly targeted in marine pollution surveillance programs [27–29]. Generally, natural metal concentrations in sediments are not detrimental to inhabiting organisms [30]. For metabolism, organisms essentially require some key micronutrients (i.e., iron, copper, zinc, manganese, cobalt) [31] but at high metals concentration can pose a toxicological threat to marine organisms [8,32,33]. Sediment features such as mineralogy, sediment texture, metal properties, pH, organic matter, oxidation–reduction potential, and adsorption and desorption processes are important controlling factors for the accumulation and availability of heavy metals in the sediment [34–42]. Thus, sediments are considered as sources of heavy metals in marine environments and play a key role in their deposition and transmission along the trophic chain and must be therefore monitored. The current European Union (EU) legislation for the protection and conservation of the marine environment emphasizes the need to evaluate and limit the concentrations of contaminants, and to undertake preventive measures to re-establish a Good Environmental Status in impacted marine areas, as requested by Marine Strategy Framework Directive (MSFD, 2008/56/EC, Commission Decision 2010/477/EU).

The present study aims to recognize the sedimentological processes that drove the formation of the present sediment bodies on the base of the geochemical and sedimentological properties of the superficial marine sediments in a central Italian Adriatic Sea coastal area. In addition, the trace element pollution in the surface sediment of this area is investigated. The level of pollution has been evaluated by the elaboration of contamination index and by comparison with the limits of the Italian legislation [43]. The comparison with these limits needs to be considered as an indication due to the different methods used to measure the element concentrations. The research has been carried out by using a statistical elaboration of the biogeochemical and sedimentological variables. The statistical approach allowed the determination of compositional facies, that is, areas with common geochemical, mineralogical, and sedimentological properties. Previous studies have investigated the sedimentological and geochemical processes and recognized the main source area of sediments as well as anthropic heavy metal inputs in different offshore areas of the Adriatic Sea [4,13,44–46] but little is known about the present study area that instead shows a specific role as consequence of the coastal morphology on the Adriatic Sea circulation [47].

#### **2. Study Area**

The Adriatic Sea is located between the Italian peninsula and the Balkan Region; it is an elongated basin (about 800 km long and 200 km wide) oriented NW/SE and is connected to the Mediterranean Sea through the Otranto Strait. The Adriatic sea is historically divided into three distinct sectors [48]: the northern sector is characterized

by a shallow continental shelf with a very gentle slope of about 0.02◦ until the isobath of 100 m; the middle sector, between the 100 m isobath and the Gargano Promontory, where a basin, called Mid-Adriatic Depression (MAD), or Jabuka Pit or Pomo Pit, occurs; this basin exhibits three sub-trenches, with maximum depths of 255 m (Western Pit), 270 m (Central Pit), and 240 m (Eastern Pit) [49]; the southern sector, between the Gargano Promontory and the Otranto Strait, which shows a maximum depth of 1260 m in the South Adriatic Depression (SAD). The hydrodynamic of the Adriatic Sea is characterized by a general anticlockwise circulation, aligned to the coasts that reaches its maximum in winter and spring, and by weak currents resulting in a series of clockwise and anti-clockwise gyres, in summer (Figure 1). In the eastern side of the basin, the current is directed northward (The East Adriatic Current: EAC). The EAC is characterized by warmer and high saline waters coming from the Levantine Basin (Levantine intermediate water, LIW), and from Ionian Basin (Ionian surface water: ISW) and in some cases from the Modified Atlantic Waters (MAW) following the pattern of the Bimodal Oscillating System (BiOS) [50–53]. The Adriatic Sea circulation is complicated by the formation of dense waters in the Northern Adriatic Sea (North Adriatic Dense Water: AdDW). These are cold and high salinity waters forming in winter in the Northern Adriatic Sea when the Bora wind blows [54,55]. The AdDW annually or biennially, generally in spring, flows southward close to the coast and reaches partly the MAD and flowing further south [56] (Figure 1).

**Figure 1.** Bathymetric map of the Adriatic Sea with the three morphological sectors (lower left corner), the general Adriatic Sea circulation (in the center) and the study area (on the upper right corner). WAC: Western Adriatic Current; EAC: Eastern Adriatic Current; NAdG: North Adriatic Gyre; MAdG: Mid Adriatic Gyre; SAdG = South Adriatic Gyre. AdDW: Adriatic Dense Water (in blue North Adriatic formation area of AdW-NAdDW). LIW: Levantine intermediate Water. BIOS: Adriatic-Ionian bimodal oscillating system [50–53].

> In this context, waters and the fine sedimentary load of the Po and Apennine rivers as well as of the local rivers flowing along the South Marche coast, tend to branch out offshore in summer and to be confined near the coast in winter [4,57]. Overall, the combined action of currents and waves induces a redistribution of bottom sediments towards the south and, to a certain extent, towards the open sea, in the Northern and Central Adriatic Sea [2,4]. The main sedimentary inputs of the northern and central sectors of the Adriatic Sea are located along the western coasts since the contributions from the Balkan rivers are limited,

due to the prevalent carbonatic composition of the rocks, and the confinement within the inshore basins parallel to the coast [58]. The sediment contribution of the Northern and Western Adriatic Sea rivers is represented by the northern Alpine rivers that contribute for 8 MT/y, the Po basin, contributing for 13 MT/y, and the Apennine rivers that contribute for 22 MT/y, for a total of 43 MT/y [59]. The catchment area of the Po River includes igneous, metamorphic and sedimentary rocks of the western and central Alps, carbonatic rocks mainly from the northern Alpine rivers, and a mixture of carbonatic and silicoclastic sediments from the Po alluvial System [42]. South of the Po River mouths the solid inputs of these rivers are composed mainly of terrigenous sediments with varying contents of carbonates [60]. The solid load of the minor rivers Southern of the Po River is estimated to be 16.9 × 106 MT/y, between the Po River and Ancona, and 7.8 × <sup>10</sup><sup>6</sup> MT/y, between Ancona and Punta Penne [61].

The current sea bottom sedimentological features of the northern and central sectors of the Adriatic Sea are the result of the last sea level rise and the following and present general Adriatic Sea hydrodynamics and riverine inputs. Indeed, the northern and central Adriatic Sea bottom is characterized by a Holocene accretionary wedge parallel to the Italian shoreline. This wedge is made up of coarse sediments near the coast, followed by finer sediments eastward. Further offshore, the Holocene accretionary wedge gives way to coarser sediments again consisting mainly of reworked sands. These sandy substrates were left in place by the last fast transgression and are due to the outcropping of partial reworked beach sands, paralic sediments and old filled valleys [58,62,63].

The study area is located in the central part of the western Adriatic Sea, in front of the southern part of the Marche Region (from the Chienti River to the Tronto River), from the shoreline to a depth of about 50 m (Figure 2). The main tributaries in the area are the Chienti, Tenna, Aso and Tronto Rivers. These rivers are located in an area with neighboring industrial activities, are contaminated with domestic and industrial wastewater and are exposed to pesticides used in surrounding agricultural areas [64]. From west to east, these rivers drain the carbonatic rocks of the Apennines, followed by the arenitic and clay Miocene rocks, the Plio-Pleistocene clays, and the arenitic and conglomeratic Pleistocene formations (Figure 2a). From a morphological point of view, the study area is characterized by a flat bottom gently sloping eastward, with weak dipping near the coast, slightly higher dipping in the center and again a less sloping to the East (Figure 2b). Only two minor river catchments, namely Ete and Tesino, do not include the Apennine carbonatic rocks (Figure 2). In the study area, the Holocene wedge is formed by an accretionary prism that tends to become thinner towards the coast and offshore [47]. Moreover, this coastal area undergoes intensive bottom sediment resuspension due to strong autumn and winter storms [65–67].

**Figure 2.** Study area with the mainland geological setting and sampling stations, red and yellow dots, respectively (**a**); topographic section along the study area (**b**). The map has been generated using QGIS 3.10.4-A Coruña. Rivers courtesy of ISPRA (http://www.sinanet.isprambiente.it/it/sia-ispra/download-mais/reticolo-idrografico/view), and the geological setting courtesy of Geoportale Nazionale (http://www.pcn.minambiente.it/viewer/).

#### **3. Materials and Methods**

#### *3.1. Sampling*

Sediment samples were collected by a box-corer (sizes 101 × 17 w × 25 d cm) following a grid of 64 stations located in front of the Southern Marche Region coast, from shoreline to offshore (Figure 2a). The sediment samples were collected during the CASE1 cruise, in April 2010, carried out by the CNR-ISMAR of Ancona on the M/N G. Dallaporta. Each boxcore was described for the macroscopic characteristics (color by Munsell tables, presence of organism and bioturbation, hydration and texture, sedimentary structures, grain-size) of the lateral and superficial surfaces. pH and Eh were measured in the sediment by punching in electrodes in the first 2 cm. The sediment samples were then collected at different depth on the base of the texture of the sediment core. On the basis of present research, only the superficial samples (0–2 cm) have been considered. Each sediment sample was split in aliquots immediately after the sub-sampling for the various analyses. The aliquot for the grain-size and water-content analyses was stored at 4 ◦C, while samples for geochemical analyses were stored at −20 ◦C and then freeze-dried.

Some samples (red dot in Figure 2a) were analyzed for grain-size and biogeochemical variables (total carbon (TC), total nitrogen (TN), organic carbon (OC), δ13C, and chemical elements); while other samples (yellow dot in Figure 2a) were analyzed only for grain-size, TC, TN, and OC.

#### *3.2. Grain-Size and Organic Geochemistry*

Grain-size analyses were carried out on wet sediment samples pre-treated with H2O2- 16 vol. solution to remove organic matter. The coarser fractions were dry-sieved over the range 2 mm down to 0.063 mm by a stack of sieves ISO 3310 in accordance with UKAS Traceability Test Sieve LAB22-1. Bioclast component (>2 mm fraction) was separated by sands (2 mm < sands > 0.063 mm), while the finer fractions were analyzed by X-ray sedigraph (Micrometrics 5000D) [45].

The TC and TN were measured with a FisonsNA2000 elemental analyzer on the freeze-dried bulk sediment, while the OC was analyzed after acidification in hydrochloric acid 1.5 M [68] by the same instrument. The inorganic carbon (IC) was determined from the difference between TC and OC.

Stable isotopic analyses of OC (δ13C) were carried out on samples by using a FINNI-GAN Delta Plus mass spectrometer directly coupled to the elemental analyzer. Stableisotope data were expressed in ‰ relative to the variation (δ) from the international PDB standard.

#### *3.3. Inorganic Geochemistry*

The analyses of chemical elements (Ag, Al, As, Ba, Be, Bi, Ca, Cd, Co, Cr, Cu, Fe, Ga, Hg, K, Mg, Li, Mn, Mo, Na, Ni, P, Pb, Sb, S, Sc, Sr, Te, Ti, Tl, U, V, W, Y, Zn, and Zr) were carried out on freeze-dried sediment sample after acid digestion employing HF, HClO4, HNO3, and HCl and the following determination of the concentration of each element in the eluate solution by Inductively Coupled Plasma-Optical Emission Spectroscopy (Activation Laboratory LTD, Ancaster, ON, Canada).

To identify possible anthropic contributions, some pollution indicators, such as the contamination factor (C*f*), the degree of contamination index (C*d*), the enrichment factor (EF), the geoaccumulation index (I*geo*), and the pollution load index (PLI) of heavy metals and other trace elements were calculated [69] (Table 1). Furthermore, element concentrations were compared to the threshold values of the Italian legislation [43], to the general abundance data reported for the marine shale [70] (Table 2) and the ecotoxicological risks with the Sediment Quality Guidelines (SQGs) indicated in Rachel and Wasserman, 2015 [71].


**Table 1.** Summary of the classification categories of sediment contamination derived from the different applied indicators.

**Table 2.** Reference values, local background, normative limits and ecotoxicological reference concentrations for the elements investigated in the present study, when available 1 [70].


Contamination factor (C*f*). The C*<sup>f</sup>* [5] is an index developed to evaluate the contamination of a given toxic substance in a basin, and it is calculated as:

$$\mathbb{C}\_{fi} = \frac{\mathbb{C}\_{ci}}{\mathbb{C}\_{bi}} \tag{1}$$

where C*ei*is the concentration of the substance *i* in sediment samples, and C*bi* is the background values of the same substance indicated in Table 2. In this work, the C*fi* was calculated for all elements even if only trace elements that may be affected by human activities (As, Ba, Co, Cr, Cu, Ni, Pb, V, and Zn) or that can be indicative of peculiar processes (Ga, Sc, Sr, Y, and Zr) have been discussed.

The background values (C*bi*) of the element *i* and of the other elements not considered for the calculation of C*fi* (Table 2) were deduced by using the concentrations in the sample collected at a depth of 105 cm of the sediment core 47 recovered in the study area (Figure 2a, green star) during the PRISMA1 project [4]. The element concentration in the core of the PRISMA1 project were determined by XRF following the method specified in [5] so that the concentrations can be compared to those determined by strong acid dissolution, both referring to the total concentrations. The core 47 is located inside the study area and the level 105 is referred to a time previous the industrial age as deduced by the radiometric and metal anthropogenic data of the PRISMA1 and by the accumulation rates determined in this area by previous studies [72] so that is not affected by anthropic inputs. Furthermore, by the normalization respect to the Al or Ti content determined in the same sample, it can be compared to samples with different grain size composition. Other methods such as those used in an area in front of the Abruzzo coast [73] can be useful applied when available only superficial samples.

Enrichment factor (EF). The EF has been developed to evaluate the metal contamination. The EF normalizes the trace element content with respect to a sample reference metal, in this case the Al:

$$\text{EF} = \frac{\frac{\text{M}}{\text{A} \text{l}} \text{S} s}{\frac{\text{M}}{\text{A} \text{l}} \text{B} s} \tag{2}$$

where (M/Al)*Ss* is the ratio of each metal and Al concentrations of the sediment sample and (M/Al)*Bs* is the same ratio in the background sample (Table 2). The ecological risks based on EF values are categorized according to Table 1.

Geoaccumulation index (I*geo*). The I*geo* is an indicator of heavy metal contamination of sediments with respect to the background natural levels (C*bi*). I*geo* is expressed for each metal as:

$$\mathbf{I}\_{\mathcal{G}\mathcal{C}\mathcal{O}} = \frac{\mathbf{C}\_{\mathcal{C}i}}{\mathbf{1}.5 \times \mathbf{C}\_{\mathcal{H}i}} \tag{3}$$

Based on the I*geo* values sediments are classified according to the classes shown in Table 1.

Modified degree of contamination index (mC*d*). The mC*d*, is the sum of all contamination factors of various heavy metals. It is obtained as follow:

$$\text{mC}\_d = \frac{\sum\_{i=1}^{i=n} \text{C}\_f}{n} \tag{4}$$

in which *n* is the number of analyzed elements and *i* = *i*th element. The classification of mC*<sup>d</sup>* is shown in Table 1. In this work, the mC*<sup>d</sup>* has been calculated considering only metals or trace elements that show almost one value of C*<sup>f</sup>* > 1 (As, Ba, Ga, Pb, Sr, and Zn).

Pollution load index (PLI). The Pollution Load Index (PLI) evaluates the heavy metal contamination of sediment samples as:

$$\text{PLI} = \left(\mathbb{C}\_{f1} \times \mathbb{C}\_{f2} \times \mathbb{C}\_{f3} \times \dots \times \mathbb{C}\_{fn}\right)^{1/n},\tag{5}$$

where C*<sup>f</sup>* is the contamination factor and *n* is the number of metals. The PLI was calculated retaining only metals or trace elements that show at least one value of C*<sup>f</sup>* > 1 (As, Ba, Ga, Pb, Sr and Zn). According to the contamination degree, the PLI data were classified as reported in Table 1 [74].

#### *3.4. Statistical Analyses*

Selected variables (i.e., major and trace element, grain-size, OC, IC, δ13C, TN, OC/TN, pH, and Eh) were processed using a multivariate statistical analysis (i.e., Factor Analysis, FA). This analysis allowed us to recognize the primary relationships between a series of samples, greatly reducing the size of the multidimensional system without losing a significant amount of information. The analysis consisted of a Q-mode FA applied to the sediment samples and their compositional variables [75]. The FA was carried out using the following steps [76]: (1) standardization of the initial data between the minimum and the maximum value for each variable; (2) correlation of the variables using the similarity coefficient cosθ [77]; (3) choice of a number of factors for which the sum of the individual variances ranges between 90–95% of the total variance of the system so that not much significant information is lost; and (4) final rotation of the axes-factors of the new system keeping them orthogonal. This last step was carried out to obtain better guidance on the original samples and to preserve the independence of the variables (Varimax criterion).

The FA was carried out to avoid an excessive number of variables respect to the number of samples. The selected variables of each sample were processed to obtain statistical factors clustering similar geochemical and sedimentological compositions. These factors represent biogeochemical-sedimentary facies (BSF), that is, the component of each sample that has been subjected to the same transport and depositional processes and whose sediments are from the same source areas. The areal mapping of the score of each factor extracted for each sample allows us to obtain the areal distribution of the different BSF and then to infer the main sedimentological and biogeochemical processed that took place in the area. The areal distribution maps of the BSF as well as of the single variables were drafted by using the tension continuous curvature spline method [78]. All the biogeochemical and sedimentological variables were plotted according to their areal distribution following a standardized method that creates a grid file with GMT v. 6.0.0 [79] and plots the map by QGIS v. 3.10.4-A Coruña. The areal distribution of each variable is shown in Figure S1 (Supplementary Materials).

#### **4. Results and Discussion**

A summary of statistics of the investigated variables is reported in Table 3. Selected chemical elements (i.e., Cd, Sb, Tl, U, W) have not been considered as their concentrations were either below the detection limits or above detection limits in a very limited number of samples (in parenthesis after the element): Bi (1), Hg (2).

Table 3 also includes information about Ag (9), Mo (16), Te (19), Be (29), and As (41) although they were not observed in some of the samples. All these elements were excluded from the multivariate statistical analysis (the complete dataset is reported in Supplemental Table S1). Table 2 presents some reference values used in the discussion, the background concentrations considered for some elaborations, regulatory values and the ecotoxicological sediment values. Meanwhile, Table 4 includes basic statistics relative to the various indexes calculated for describing the sediment status.


**Table 3.** Summary statistics for the investigated variables.


**Table 4.** Statistical values (mean, median, min, max and dev.st.) of the contamination indexes data and pollution indicators.

#### *4.1. Sedimentological Processes and Facies Recognition*

The statistical elaboration of the biogeochemical and sedimentological variables highlights the presence of three main factors explaining, on overall, 94.6% of the total variance. These factors mainly represent the three key BSFs in the coastal marine area of the Southern Marche.

The most significant *facies* (BSF1) explains 79.2% of the total variance and is characterized by sediments with (Figure 3a) high contents of clay and fine silt, some elements (i.e., Al, Ti, Mg, K, Na, P, Fe, Co, Cu, Zn, Ni, Pb, V, Ga, Li, Y, and Zr mainly linked to clay minerals), OC that is directly linked to the Organic Matter (OM), and TN that is contained both in the OM and in the clay minerals. The SBF1 is therefore sedimentological facies consisting primarily of fine pelitic sediments with abundant clay minerals and OM. The

extracted scores of these facies increase seaward, starting from values close to zero, that imply absence of pelitic sediment, clay minerals and OM, near coast, to values close to 1 (i.e., sediments made up of almost all fine sediments) (Figure 3b). By considering the areal distribution of the fine pelitic, organic and clay-rich sediments of BSF1 and the general cyclonic circulation of the Adriatic Sea as well as the reworking and removing action of the wave near the coast, the sediments of the BSF1 are inferred to be mainly sourced from the fine load of the Po River as well as from that of the Apennine rivers in the northern part of the study area. For this reason, the BSF1 facies can be assimilated and named as the *Padanic Facies* of Spagnoli et al. (2014) [4]. The fine particles introduced in the Adriatic Sea in front of the river mouths and are then resuspended by the wave action and transported southwards by the WAC [2]. As a result, the fine particles settle at greater depths in a belt area aligned to the coastline, where the action of the wave motion on the seabed is weaker [4]. Furthermore, the WAC confines the suspended sediments near the coast, preventing their spreading eastward limiting their sedimentation towards the center of the basin. The fine sediments of the BSF1 are subject to a continuous reworking process acting on the sea bottom. This consists in repeated resuspension and redeposition processes, due to strong storms that result in the transportation and dilution processes of the BSF1 sediments southward. As a consequence of this sedimentological and hydrographic set up, the present offshore sediment accumulation is very feeble, and the relict sediments connected to the last sea level rise tend to surface in the deeper side of the north-central Adriatic Sea. The BSF1 facies approaches towards the coast between the Chienti and the Ete Vivo rivers and in front of the Tronto river where it also shows a maximum (Figure 3b). This distribution near the coast suggests that the Chienti, Tenna and, to a greater extent, the Tronto rivers discharge more fine sediment than the other rivers in the southern Marche.

(**b**)

**Figure 3.** Histogram of the factor scores (**a**) and map of the areal distribution of the of the BSF1 (**b**).

The second facies (BSF2) represents 14.5% of the total variance and is marked by high contents of sand, IC, Ca, Sr, OC/TN, pH, and Eh (Figure 4a). Calcium and IC are typical of sediments with abundant carbonate, while high values of pH and Eh are associated to coarse grain-sizes. Based on these characteristics, it can be inferred that the BSF2 is mainly made up of a coarse sandy sediment enriched in carbonates. The areal distribution pattern of the BSF2 (Figure 4b) is complementary to that of the BSF1, with values that suddenly increase the near the coast, where they show values close to 1, meaning sediments consisting almost exclusively of carbonate sandy materials and decrease offshore where they reach value lower than 0.1. The origin of the BSF2 sediments is ascribed to the coarse sandy contributions of the local rivers of the Marche Region. They drain the carbonate rocks of the central Apennine and discharge their coarse sediments near the coast, while finer sediments are removed by wave action. The BSF2 facies can be named as *Coastal Facies* being mainly affected by coastal processes and inputs. The BSF2 presents a smaller extension towards the offshore in front of the Chienti and Tenna rivers and, to a greater extent, in front of the Tronto River where, there is also an area marked by the lowest values. This trend in the areal distribution of BSF2 confirms the greater inputs of fine material from the Chienti, Tenna and, to a greater extent, from the Tronto rivers.

**Figure 4.** Histogram of the factor scores (**a**) and map of the areal distribution of the of the BSF2 (**b**).

The third facie*s* (BSF3) explains 1.9% of the total variance and is defined by abundant silts, P, Mn, Ba, Sr, and partially Zr (Figure 5a). The highest values of the BSF3 are mostly found along a belt parallel to the coast in an intermediate position between the BSF1 and BSF2 maxima (Figure 5b) with three maximum areas. By considering the distribution of the BSF3 sediments in relation with the hydrography of the central Adriatic Sea, the silty sediments of the BSF3 are the results of the accumulation at depth where the wave action can remove the finer granulometric fractions, but where the coarse sandy coastal material supplied by the local rivers and reworked by the wave motion, is not deposited. Furthermore, the presence Zr and P also suggest the presence of heavy minerals like zircon and apatite, while the occurrence of Sr and Mn as well as IC suggest the presence of carbonates. These minerals support the residual nature of these sediments in which heavy minerals and carbonate concentrations increase for the removing of lighter minerals and finer grains. The characteristics of the BSF3 are similar to the *Residual Facies* of Spagnoli et al. (2014) [4] and can then be partially considered as the same *Residual Facies*. The two minima near the coast are in front of the Ete Vivo and Tesino rivers: the two rivers that do not drain the limestone rocks of the Sibillini mountains. This means that the residual sediments are partially associated with the carbonate rocks.

**Figure 5.** Histogram of the factor scores (**a**) and map of the areal distribution of the of the BSF3 (**b**).

By considering single elements, it is worthy to note that the elements that present high affinity with the clay minerals (i.e., Al, Co, Cr, Cu, Fe, Ga, K, Li, Mg, Mn, Na, Ni, P, Pb, Sc, Ti, V, Y, Zn, Zr) show a distribution pattern very similar to that of the BSF1

*facies* (*Padanic Facies*) (Figures 6–8) because they naturally tend to accumulate with finer fractions. On the other hand, elements having high affinity with the carbonate minerals (i.e., Ca and Sr) show distribution patterns (Figures 6f and 7e) similar to that of the *Coastal Facies* BSF2 (Ca) and BSF3 (Sr). This is in full agreement with the findings of the statistical elaboration that connects the same elements with the clay minerals and with the carbonate minerals. For this reason, the anomalies regarding these patterns have been considered to infer anthropic inputs or local natural processes. In this context, an anomaly is recorded in front of the Tronto River where a siliciclastic fine sediment enriched in S is recorded (Figure 6). This is also supported by the sediment description of the box-core upon collection that consisted of compact gray mud with blackish veins and burrows. Following the bathymetric reliefs that delineate a depression, the geochemical composition of this area is related to a strong erosion of the bottom that allows the outcropping of older, weak diagenized siliciclastic sediments.

**Figure 6.** *Cont.*

**Figure 6.** Areal distribution of the elements in surface sediments. (**a**) Al, (**b**) Fe, (**c**) K, (**d**) Mg, (**e**) S, (**f**) Ca, (**g**) Ti, (**h**) P, (**i**) Na, (**j**) Mn, and (**k**) Li.

**Figure 7.** *Cont.*

**Figure 7.** Areal distribution of the elements in surface sediments with C*<sup>f</sup>* > 1 in some samples. (**a**) As, (**b**) Ba, (**c**) Ga, (**d**) Pb, (**e**) Sr, and (**f**) Zn.

**Figure 8.** *Cont.*

**Figure 8.** Areal distribution of the elements in surface sediments with C*<sup>f</sup>* < 1 in all samples. (**a**) Co, (**b**) Cr, (**c**) Cu, (**d**) Ni, (**e**) Sc, (**f**) V, (**g**) Y, and (**h**) Zr.

#### *4.2. Evaluation of the Contamination and Ecotoxicological Risks*

#### 4.2.1. Contamination Assessment

Only elements with Cf higher than one in some samples are herein considered.

As. Arsenic mean values are in the order of 9 ppm (Table 3) that are slightly lower than the average marine shale composition (Table 2**)** and are in line with the local background value **(**Table 2) of the southern Marche Holocene pelitic wedge (10 ppm at −105 cm). The spatial distribution indicates the occurrence of relatively high concentrations (up to 31 ppm) in the northern area (Figure 7a).These data suggest that this area has, on average, high As values likely related to the composition of the source rocks coming from North [4]. Arsenic was included in the multivariate analysis but shows a good affinity with elements of the BSF1. In two sandy areas located inshore in front of the Chienti and Ete Vivo river mouths, as well as in a pelitic belt directed north-south, in front of the Chienti River (Figure 7a), the As values are equal to or higher than the Italian thresholds (Table 2), its C*<sup>f</sup>* is higher than 1 and the EF and I*geo* fall in the second class (Table 4).The mainly sandy sites, located inshore, in front of the Chienti River (station 101) and of the Ete Vivo River (station 303) fall in the moderate enrichment class (2–5) for EF index and unpolluted to moderately polluted (0–1) for I*geo*, respectively. Furthermore, the concentration of As at station 303 is over the second threshold of the Italian legislation (20 ppm, Table 2). In addition, the station 106, in the pelitic belt in front of the Chienti River, is moderately polluted (1–2) for I*geo* and shows a concentration over the second threshold of the Italian legislation (20 ppm, Table 2). Unpolluted to moderately polluted (0–1) I*geo* values are also recorded offshore and to the Southeast of the study area, in correspondence with mainly clayey sediment, but in this case the high values are due to the inputs of fine materials coming from North, that however influence also other proximal areas, further south [73]. The As pollution in the north-western area could be ascribed to local discharges of waste of fertilizer productions along the southern Marche coast in the past decades.

Ba. The average concentrations of barium are around 270 ppm (Table 3) reaching a maximum of 329 ppm, far below the average marine shale composition and in line with local background (Table 2). The highest values of Ba are found near the coast between the Chienti and Aso Rivers and in a belt along the central area that mainly corresponds to the BSF3 (Figure 7b). In some stations of these areas, the EF of Ba slightly exceeds 2, resulting in the moderate enrichment class, while the absolute concentrations (average 272.8 ppm; Table 2) are higher than the background value of 295 ppm in the high Al content stations (Table 2). The high Ba concentrations near the coast and in the BSF3 *facies*, particularly in the northern side of the study area, suggest a provenance of Ba from North. It also suggests a scavenging process, due to high specific weight of the barium sulphate, a heavy mineral that tends to remain in situ after reworking processes. The barium sulphate is a component of the drilling muds so that the high Ba concentrations could be due to discharges into the sea in the past of these drilling muds from hydrocarbon platform perforation as reported for other areas [2,46,80].

Ga. The values of Ga in the whole area (average 12.3 ppm, Table 3) with a maximum of 22 ppm are in line with average marine shale and with local background concentrations (16.2 ppm, Table 2) even if some stations show values of C*<sup>f</sup>* slightly above 1. The EF and I*geo* indexes are constantly lower than the minimum threshold values. Its distribution describes an almost regular increase seaward (Figure 7c).

Pb. The average values of Pb in the whole area (12.1 ppm, Table 3), with a maximum of 24 ppm, match well with the background data (18 ppm, Table 2) and the average marine shale. Some samples, however, show a C*<sup>f</sup>* slightly over 1 (Table 4). The ER and I*geo* indexes are permanently less than the minimum threshold values (Table 4). Both the present average Pb contents and the background values are strongly lower than the L1 threshold level of 30 ppm (Table 2) suggesting limited anthropogenic Pb inputs.

Sr. Strontium has an average concentration of 353 ppm (maximum of 436 ppm, Table 3) that are comparable to the local background and slightly higher than average marine shale (Table 2). This element has a strong affinity for carbonates and its concentrations show higher values in correspondence of the residual facies BSF3. Among the sedimentary quality index, the I*geo* is constantly lower than minimum threshold value, but the EF index is higher than 2 (low enrichment, Table 4) in many samples, likely due for the way the index is calculated on a normalized base.

Zn. Zinc concentration records an average of 49 ppm and a maximum of 95 ppm (Table 3) that are, in general, slightly lower than local background and average marine shale values (Table 2). The sediment quality index displays few observations of the C*<sup>f</sup>* value of 1, whereas the EF and I*geo* indexes are constantly lower than minimum threshold values (Table 4). All the observations are below the L1 threshold levels (100 ppm, Table 2) pointing therefore to a limited anthropogenic contribution for this element.

As regard the cumulative impact of the elements it is important to recall that the mC*<sup>d</sup>* and the PLI were calculated considering only metals or trace elements that show almost one value of C*<sup>f</sup>* > 1 (As, Ba, Ga, Pb, Sr, and Zn). Despite the cumulative impact of more anthropogenic substances, the values of mC*<sup>d</sup>* are always below the threshold of 1.5, so all sample sites fall in the very low polluted class (Table 4). On the other hand, the PLI shows values slightly higher than 1 in front of the Civitanova Marche, at the higher depths towards the offshore, particularly on the south-eastern part of the study area, and in front of the Tronto River. While the high values of PLI present in the offshore are attributed to the fine grain size sediments coming from the north and belonging to the BSF1 Facies (Table 3, Figure 9), those occurring in the area between Chienti and Ete Vivo Rivers could be the results of local pollution inputs, in particular the pollution in this area could have been higher if it were related to finer grain-size. In this coastal area, the pollution is due to high values of As, followed by Pb and Zn. The high concentrations of heavy metals, mainly As, could be caused by the discharge of processing residues of fertilizer production industries that were operative along the southern Marche coast in the last century.

**Figure 9.** Areal distribution of the PLI values.

The following elements exhibit a Cf lower than one.

Co. The concentrations of Co in the whole area display an average value of 7.8 ppm and reach a maximum of 13 ppm (Table 3), in general below the local background and the average marine shale (Table 2), that suggest a regional depletion compared to global references. Its distribution displays an almost regular increase seaward (Figure 8a), common to all the elements included in the BSF1. The C*<sup>f</sup>* values are constantly less than 1.

Cr. Chromium average concentration is 61.5 ppm with a maximum of 101 ppm (Table 3). The average marine shale (90 ppm Cr, Table 2) and the local background (128 ppm Cr, Table 2) values are comparable or higher indicating that there is no enrichment respect to the pre-industrial period. In some stations close to the coast, mainly sandy, some higher values are recorded and could be either related to possible anthropic inputs or to selective enrichment in heavy minerals, given the high values of other elements such as Zr and Ti. Many of the observations are above the L1 threshold level of the Italian legislation (50 ppm, Table 2). It should however be considered that, both for Cr and all the other considered elements, the values of the national legislation refer to an aqua regia digestion and not to a multi-acid digestion including HF, as the one applied in the present study.

Cu. Copper concentrations have a mean value of 14.4 ppm and a maximum value of 32 ppm Cu (Table 3), the latter is lower than the average marine shale (45 ppm) and the local background (40 ppm) ones. (Table 2). This element does not, therefore, represent an important issue for this area. This is further confirmed by the C*<sup>f</sup>* values of Cu (Table 4) constantly lower than 1 as well as by the values of the EF and I*geo* indexes, permanently lower than the minimum threshold values (Table 4). Considering this, there has not been any enrichment with respect to the pre-industrial periods. Indeed, the weak decrease in the last decades could be related to the reduction of northern riverine inputs.

Ni. Nickel has an average concentration of 32.7 ppm and a maximum of 63 ppm, which are both lower than the average marine shale and the local background. Despite of it, many stations show values above the L1 threshold level of the Italian legislation (30 ppm, Table 2). The comparison between the average values and the background data in the study area suggests that L1 threshold level of the Ni could be underestimated for this area, likely due to the different analytical method. A support is given by the C*<sup>f</sup>* values always below 1 and by the EF and I*geo* indexes constantly lower than the minimum threshold values (Table 4).

Sc. Scandium has an average concentration of 6.8 ppm and a maximum of 12 ppm (Table 3) against an average marine shale concentration of 13 ppm and a local background of 29 ppm (Table 2). It is not an element considered by the legislative side, and its C*<sup>f</sup>* values are constantly less than 1 as well as the EF ER and I*geo* indexes are permanently lower than minimum threshold values (Table 4). These data indicate that there is not Sc enrichments respect to the pre-industrial period and its decrease in the last decades could reflect the decrease of northern riverine inputs.

V. Vanadium has an average concentration of 49 ppm and a maximum value of 105 ppm. Its spatial distribution displays a clear seaward increase (Figure 8f). The values are below the average marine shale and the local background values (Table 2), likely for the diluting effect of the carbonate fraction and the relative importance of sandy sediments, both causing a general depletion of V. The C*<sup>f</sup>* value are always lower and the EF and I*geo* indexes and lower than the minimum threshold values (Table 4). This indicates that there are no enrichments of V respect to the pre-industrial period but an overall decrease. These decreases could be due to the decrease of northern riverine inputs in the last decades.

Y. Yttrium concentrations have an average concentration of 13.7 ppm and a maximum of 16 ppm, all values are well below the global reference and the local background (Table 2). The dilution effect of carbonate material is the likely explanation for this difference. The C*<sup>f</sup>* values are always lower than 1; similarly, EF and I*geo* indexes are constantly lower than the minimum threshold values (Table 4). All these evidence points to a lack of t enrichments of this element with respect to the pre-industrial period.

Zr. Zirconium has average value of 35 ppm Zr and a maximum value of 65 ppm that are well below the average marine shale (160 ppm, Table 2) and the local background (97 ppm, Table 2). The values of C*<sup>f</sup>* value are always <1, and EF and I*geo* indexes are permanently lower than minimum threshold values (Table 4). This decrease respect to the pre-industrial period is attributed to the decrease of northern riverine inputs in the last decades.

The analysis of the core 47 also allows us to establish the background values of other elements other than those considered for the contamination assessment (Table 2). An important finding arises for the comparison between the present concentrations and the background values. The concentrations of most of the elements with a high affinity to clay and silicate minerals show a decrease respect to the background values (Table 5a), while other elements with an affinity with carbonate minerals exhibit an increasing respect to the past (Table 5b). These trends suggest a change in the sedimentation pattern in the area in the last decades. Specifically, this change consists in a percentage decrease of the clay contributions coming from the North and a percentage increase in the carbonate fraction due to the local river inputs. This change could be ascribed to the lower fine solid inputs of Po River as a result of the regimentation works and the lower quantity of water flowing of the Po River due to both the pumping water for zootechnical, agricultural and industrial uses and for the climate changes in the last decades [81].

#### 4.2.2. Ecotoxicology of Central Western Adriatic Sea Bottom Sediments

The Numerical Sediment Quality Guidelines (SQGs) developed for marine and estuarine ecosystems as suggested by Rachel and Wasserman (2015) [71], are used here to evaluate the ecotoxicological state of the bottom sediments. According to Rachel and Wasserman SQGs (2015) [71], the following thresholds can be considered (Table 2): the TEL (Threshold Effect Level) that is the concentrations below which sediment-dwelling organism do not exhibit toxic effects; the ERL (effect range low) that is the concentrations below which sediment-dwelling organism rarely exhibit toxic effects; the PEL (Probable effect level) and the ERM (Effect Range Median) that are the concentrations above which toxic effects are detected frequently [82].


**Table 5.** Elements that decrease respect to the past (a); elements that increase respect to the past (b); elements for which only the background value is available (c).

Most of the samples show As values exceeding the TEL; this is the consequence of the background values of As that is rather high in this area and due to the inputs of sediments coming from the North that present high As source catchment areas. Furthermore, two samples (303 and 106 sites) show As values exceeding the PEL. In this case, as previously noted for the contamination indexes, the exceeding of the SQLs thresholds is probably due to anthropic causes due to the discharges into the sea of waste of the fertilizer production of firms that have been operative along the Southern Marche coast during the 20th century. As regards the Cr, almost all samples exceed the TEL, 11 samples exceed the ERL and 3 samples overcame the PEL. The high contents of Cr that generally exceed the lower levels of some SQGs are ascribed to the composition of the sediments coming from the North that exhibits high Cr levels [14] and are a well-known geochemical signature sediment associated to the Po river in northern areas [83–85]. As regard the Ni, 34 stations exceed the TEL, 25 the ERL, and 16 the PEL. In this case, as for the Cr, these excesses could be ascribed to the composition of the sediments coming from the North, which naturally show high Ni levels, or due to the sediments affected by anthropogenic input in the North Adriatic. Finally, in the cases of Cu, Pb e Zn no station presents contents higher than the TEL.

#### **5. Conclusions**

The present work analyzes the bottom sediments in front of the Marche Region by using a multi-proxies and integrated approach. This allows us to recognize and to map the distribution of three main sedimentological and geochemical facies: the *Padanic Facies* (BSF1), the *Coastal Facies* (BSF2) and the *Residual Facies* (BSF3). The Padanic BSF1 *Facies* is characterized by fine siliciclastic sediments coming from Northern areas, reflecting the regional sediment dispersal pattern, mostly originating from the Po River, and transported southwards by the WAC and towards offshore by wave motion of the strongest storms. The BSF2 *Coastal Facies* is mainly defined by coarser sediments enriched in carbonatic minerals, coming from local rivers and partly mixed, by the wave motion of the stronger storms, with the sediments of the BSF1 *Facies*. The BSF3 *Residual Facies* is mainly composed by silty sediments coming from local rivers that are partially sorted from the clayey from the wave motion. Towards the offshore, this facies mixes progressively with sediments of Padanic sources coming from the North. The distribution of these three facies is the result of two principal processes. The first is the predominance of sedimentary inputs coming from the North and from local rivers, hence, the prevalence of sediments coming from source areas with different mineral-petrographic composition. The second regards the general hydrodynamic of the Adriatic Sea such as the cyclonic circulation that mainly flows southward near the Italian coast and the wave motion of the stronger storms directed perpendicularly to the coast.

The data processing of trace elements and heavy metals concentrations also allow us to determine the background reference values for the study area. It is noteworthy to stress that the Cr and Ni values recorded in the past and in the surface sediments of this study area are above the L1 threshold levels of the Italian legislation (DM 173/2016). On the other hand, the present and past concentrations of Pb are generally lower than it. This suggests a possible underestimation of Cr and Ni thresholds and an overestimation of the Pb threshold. Therefore, a further study in order to validate or reject these discrepancies on Cr, Ni and Pb reference values should be carried out.

The present study also identifies a small coastal area, between the Chieti and Ete Vivo rivers, that is probably affected by an As and Ba local pollution. The high contents of As could be ascribed to local discharges of waste of fertilizer productions during the 20th century. Moreover, the pollution of Ba is probably related to past discharges into the sea of drilling muds coming from oil platform perforation as reported for other area on the Central Adriatic Sea. Finally, the comparison between the background concentrations and the current average concentrations point to a general decrease in Al, Ti, P, Co, Cr, Cu, Ga, Ni, Pb, Sc, V, Y, and Zn and an increase in Ca and Sr. These trends suggest a change in the sedimentation in the last decades characterized by a decrease in the clayey contributions coming from the north and an increase in the carbonate fraction coming from the local rivers. This could be ascribed to the lower solid inputs of the Po River as a result of the regimentation works and the lower water flow prompted by both climate changes and pumping for irrigation, farm, civil and industrial uses.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/2076-3 417/11/3/1118/s1. Figure S1: Areal distribution of the elements in surface sediments. Ag (a), Mo (b), Be (c) and Te (d); Table S1: Values of the biogeochemical and sedimentological data and pollution indicators.

**Author Contributions:** Conceptualization, F.S.; Data curation, F.S., R.D.M., E.D., E.F., F.F. and P.G.; Formal analysis, F.S., E.D., E.F., F.F. and P.G.; Funding acquisition, F.S.; Investigation, F.S., F.F. and P.G.; Methodology, F.S., R.D.M., E.D., F.F. and P.G.; Project administration, F.S.; Resources, F.S.; Software, F.S., R.D.M. and P.G.; Supervision, F.S.; Validation, F.S., E.D., F.F. and P.G.; Visualization, F.S.; Writing—original draft, F.S., R.D.M., E.D., E.F., F.F. and P.G.; Writing—review & editing, F.S., R.D.M., E.D., E.F., F.F. and P.G. All authors will be informed about each step of manuscript processing including submission, revision, revision reminder, etc. via emails from our system or assigned Assistant Editor. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study has been conducted with the financial support of PERSEUS Project, grant number 287600

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data is contained within the article and the Supplementary Materials.

**Acknowledgments:** We want to give our thanks to the crew of the M/N Dallaporta, to the Giuseppe Corti and the Stefania Cocco of the Università Politecnica delle Marche, to Fabio Zaffagnini, Laura Borgognoni, Eva Turicchia and Piero Ferracuti, to Giuseppe Caccamo and Elisa Ghetti of the IRBIM Ancona, for their valuable help given during the sampling, analysis and data processing phases.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


MDPI St. Alban-Anlage 66 4052 Basel Switzerland www.mdpi.com

*Applied Sciences* Editorial Office E-mail: applsci@mdpi.com www.mdpi.com/journal/applsci

Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Academic Open Access Publishing

mdpi.com ISBN 978-3-0365-8817-9