Geochemical and Radiometric Assessment of Romanian Black Sea Shelf Waters and Sediments: Implications for Anthropogenic Influence

Abstract

The Northwestern Black Sea shelf, strongly influenced by Danube discharge and coastal activities, provides an effective setting for separating lithogenic controls from localized anthropogenic inputs. We applied a multi-proxy geochemical–radiometric approach to Romanian shelf waters and surface sediments. A CTD–Rosette was used to quantify nutrients, chlorophyll-a, TOC, and TN. Dissolved metals and PAHs were measured in seawater, while surface sediments were analyzed for CaCO₃, TOC, trace metals, and γ-emitting radionuclides. Multivariate statistics (PCA/FA) were used to resolve the dominant environmental controls. Summer stratification was characterized by the bottom-layer maxima of PO₄³⁻, SiO₄⁴⁻, and NH₄⁺ and a pronounced subsurface chlorophyll-a maximum at 12–16 m. Surface-water Σ16PAH ranged from 134 to 347 ng L⁻¹ and was dominated by low-molecular-weight compounds, with episodic nearshore enrichment in high-molecular-weight species. In sediments, CaCO₃ ranged from 7.6 to 29.9% and TOC from 0.11 to 0.96%. Trace metals were generally low. Pb and Hg peaked at nearshore station S23, whereas mean Ni (38.88 ppm) slightly exceeded the 35 ppm guideline, consistent with natural Fe/Mn-oxide association. PCA/FA identified a terrigenous axis (Fe-Al-Ti-V-Ni-Cr), a carbonate axis (CaCO₃; Sr where available), and an anthropogenic factor (Pb, Hg, HMW-PAHs). γ-spectrometry provided a compatible radiometric baseline that supports the multi-proxy interpretation.

1. Introduction

Marine sediments are highly fractionated crustal materials supplied to the ocean from multiple sources [1]. In nearshore settings, river-borne detritus often dominates, so sediment composition largely reflects catchment lithology and hydrodynamic sorting. At the same time, sediments efficiently accumulate trace metals and organic carbon associated with human activities, particularly near densely populated and industrialized coasts [2]. Because sediments integrate inputs over time and can act as secondary sources under physical disturbance or redox changes, they are central to environmental assessment and to ecosystem-based management frameworks, such as the Marine Strategy Framework Directive (MSFD) [3], which requires evaluation of hazardous substances and seafloor integrity in support of Good Environmental Status. Globally, interpreting trace-metal and organic-contaminant distributions requires accounting for natural “carrier” phases and facies-dependent dilution. Trace elements typically partition to the fine fraction, aluminosilicates, and Fe/Mn oxides/hydroxides, whereas carbonate content can dilute terrigenous components and shift apparent concentrations independent of source strength. Such matrix controls complicate source attribution in many coastal seas, where port activities, urban runoff, industrial discharges, and maritime traffic may generate localized hotspots superimposed on strong natural gradients. Consequently, multivariate tools (e.g., PCA/FA) are widely applied to separate the dominant geochemical controls and infer the source/transport pathways, but robust attribution benefits from combining complementary tracers rather than relying on a single chemical group. The Black Sea provides an instructive case because it is a semi-enclosed basin with restricted exchange and strong riverine forcing, which makes its coastal zone particularly sensitive to both natural variability and anthropogenic pressure. On the northwestern shelf, hydrography, biogeochemistry, and sedimentation are strongly modulated by the Danube, the largest freshwater source to the basin, whose nutrient loads and stoichiometry have driven long-term eutrophication dynamics and ecosystem change along the Romanian coast [4,5]. Modeling and observational syntheses indicate that not only nutrient magnitude but also nutrient balance shapes primary production and food-web structure, with phosphate often emerging as a key control during the peak eutrophication period and its partial reversal [6]. This hydro-biogeochemical template is coupled to a lithofacies framework: nearshore siliciclastic inputs transition to more carbonate-rich sediments offshore. Classical regional classifications distinguish non-carbonate (<10% CaCO₃), slightly calcareous (10–30%), carbonate (50–70%), and highly carbonate (>70%) surface sediments, underscoring facies-driven dilution/enrichment as a first-order control on trace-element and organic matter patterns [7]. Along the Romanian shelf, recent assessments generally report low-to-moderate dissolved and sedimentary metals and polycyclic aromatic hydrocarbons (PAHs), with spatial heterogeneity and localized enhancements near ports and urban areas [8]. PAH source apportionment commonly uses compositional diagnostics (e.g., LMW/HMW contributions and ratios such as Ant/(Ant + Phe), Fl/(Fl + Py), and BaA/(BaA + Chry)), which can help distinguish petrogenic inputs (fuel/oil handling, maritime traffic) from pyrogenic inputs (combustion) but require caution because weathering and transformation can alter diagnostic ratios [9]. Despite the growing body of regional work, many studies remain matrix- or pollutant-group specific, limiting the quantitative evaluation, within a single statistical framework, of how lithogenic carriers and facies effects interact with localized anthropogenic inputs across multiple contaminant classes [10,11,12,13,14,15,16]. This fragmentation constrains the ability to identify emerging hotspots across pollutant classes, to interpret concentrations relative to natural background and dilution controls, and to support MSFD-oriented, multiparameter status evaluation and trend detection. Multi-proxy strategies that integrate elemental, organic, and radiometric indicators can address these limitations by providing orthogonal constraints on sources, carriers, and depositional context [17]. In practice, high-sensitivity elemental methods (e.g., ICP-MS) complement non-destructive X-ray fluorescence (XRF/EDXRF) for major and trace metals; targeted chromatographic methods resolve compound-specific PAHs; and high-resolution γ-ray spectrometry on sediments constrains inventories of natural-series radionuclides and anthropogenic markers, offering an independent radiological signature of sediment matrices and depositional environments [18]. Best practice for environmental γ-ray applications is established in IAEA guidance and interlaboratory exercises. Consistent with these recommendations, γ-assays are most reliable in solid matrices where detection limits are compatible with environmental activities [19]. From a management perspective, interpretation of chemical status also requires harmonized benchmarking against the Romanian regulatory framework (Order 161/2006) [20], which operationalizes the EU Water Framework Directive for surface waters and is routinely used in regional status assessments [20]. Against this background, the present study uses the Romanian Northwestern Black Sea shelf as a case study to quantify, within a unified multi-proxy and multivariate framework; how natural sedimentary matrices versus localized human pressures structure co-variations among hydro-biogeochemical indicators in the water column (nutrients, chlorophyll-a, and TOC/TN), dissolved contaminants (metals, PAHs), and sediment facies and inventories (CaCO₃, TOC, major/trace elements, and Hg) together with sediment γ-ray signatures. By integrating these proxies and applying PCA/FA, we aim to provide a process-aware baseline that separates lithofacies effects from spatially restricted anthropogenic signals, thereby supporting environmental status assessment and future trend detection on the Romanian shelf.

2. Materials and Methods

2.1. Study Area

The investigated sector lies along the Southern Romanian coast between Cape Agigea (N) and Cape Tuzla (S), encompassing the nearshore zones of the Eforie North–Costinești resorts and cross-shelf transect over the inner shelf to the 20–30 m isobaths (Figure 1). The station network is part of the long-term monitoring program coordinated by NIRD GeoEcoMar and covers transitional and coastal waters, as well as the inner shelf. Several stations overlap the Natura 2000 sites ROSCI0197 and ROSCI0273, ensuring relevance for protected-area management. Geomorphologically, the shoreline is dominated by high coastal cliffs (approx. 20–30 m) developed in Sarmatian limestones draped by a loess mantle, locally interrupted by coastal barriers and beaches 20–150 m wide. Between the sea and Lake Techirghiol, a barrier–lagoon system has historically acted as a sediment transit and storage zone. At present, its morphodynamic balance is influenced by major coastal works (Constanta/Agigea port jetties) and a regional sediment deficit. The sector belongs to the southern coastal unit, subdivided into relatively autonomous sediment cells. Between Agigea and Tuzla, bioclastic inputs (mollusk-shell fragments) dominate on the Neogene limestone platform, which is veneered by a thin sand cover—controls that govern cross-shore sorting and beach mobility [7]. The hydro-physical regime is characterized by predominantly southward longshore currents, modulated by N-NE-E winds and the counterclockwise Black Sea Rim Current. Surface salinity varies seasonally (minimum in spring, maximum in summer) under Danube discharge influence, and the summer thermohaline structure typically exhibits a thermocline shoaling to approx. 15 m near the coast [21,22]. Wave height and propagation direction depend strongly on wind forcing and on diffraction/reflectance induced by jetties, shaping nearshore energy distribution and longshore transport. These hydrodynamic features regulate biogeochemical fluxes (nutrient cycling, resuspension) and provide the physical context for the geochemical and radiometric assessments reported here [23,24].

Ecologically, the Eforie-Tuzla sector comprises a mosaic of habitats (soft-rock/cliff coasts, sandy beaches, bioclastic platform, and the Techirghiol lagoonal system) with essential roles in biogeochemical processes and biodiversity support [25,26]. Benthic communities on the Northwestern Black Sea shelf remain vulnerable to multiple pressures, including eutrophication, hypoxia, coastal development, bottom fishing, and coastal protection works, that can alter ecosystem structure and function [27,28,29,30,31,32,33]. Despite recent reductions in some chemical pressures along the Romanian sector [34,35], the proximity of ports and urbanized shorelines maintains potential hotspots for metal and organic contaminants, justifying the multi-proxy approach adopted here.

2.2. Sampling

A coordinated field campaign was carried out between 19 and 27 May 2025 along the Southern Romanian Black Sea coast (Eforie Nord–Eforie Sud–Cape Tuzla–Costinești), covering the inner continental shelf to ~30 m water depth (Figure 1; 20 m and 30 m isobaths). The work was undertaken within the Program Nucleu project PN 23.30.02.02, whose core objective is the systematic mapping of benthic habitats in the onshore and inner-shelf sector of the Romanian coastline. Under this framework, similar surveys are organized annually, providing the institutional basis for progressive, multi-year data accumulation and for future cross-seasonal comparisons, even though the present study reports result from a single early-summer (late-spring) campaign. Sampling stations were not selected as a simple linear transect, nor were they necessarily visited in the numerical order shown in Figure 1. Instead, station selection followed a substrate-driven design: candidate areas were first surveyed using side-scan sonar, and stations were then positioned to capture the dominant seabed types and facies transitions (e.g., sandy and sandy-mud substrates, finer muddy deposits, and locally biogenic accumulations), ensuring that collected samples are representative of the main habitat and sedimentary units targeted by the benthic habitat mapping program. Consequently, the sampling route was optimized operationally (weather window, sea state, vessel logistics, and safe access), and the station numbering in the figure serves as an identifier, not as a strict chronological sequence of sampling.

2.3. Hydrographic Profiling and Water Collection

A total of 25 surface-sediment samples (0–5 cm) were collected at fixed stations (S01–S25). Hydrographic profiling and seawater sampling were conducted at six stations using a CTD–rosette system comprising a Sea-Bird CTD (SBE series) coupled to an SBE 32 carousel equipped with twelve 5 L Niskin bottles (Figure 2). Continuous downcast and upcast profiles were acquired for pressure, temperature, conductivity/salinity, dissolved oxygen, fluorescence, and turbidity. The CTD package included a temperature sensor (SBE 3F), a conductivity sensor (SBE 4C), a dissolved oxygen sensor (SBE 43 membrane polarographic detector; accuracy ± 8 µmol kg⁻¹), and a fluorescence/turbidity sensor (ECO FLNTU; chlorophyll-a detection limit 0.015 µg L⁻¹). Fluorescence was calibrated following manufacturer specifications and converted to chlorophyll-a concentrations (µg L⁻¹). CTD data were acquired and processed using standard Sea-Bird software (SeaSave v7; SBE Data Processing Win32 v7.26.7). Profiles were quality-controlled and then binned and averaged at 1 dbar resolution. Sampling depths for the Niskin bottle collection were selected in real time from the downcast profiles to target key biogeochemical interfaces. Rosette-collected water was analyzed for dissolved inorganic nutrients, chlorophyll-a, total organic carbon (TOC), total nitrogen (TN), dissolved metals (including Hg), and dissolved polycyclic aromatic hydrocarbons (PAHs). Because the primary objective of the MN 270 survey was to resolve the spatial heterogeneity of seabed facies and sediment geochemistry across the nearshore–inner shelf, sediment sampling was implemented in a comparatively dense network (S01–S25). In contrast, full-depth CTD–rosette operations and the associated water-column analytical suite are substantially more time- and resource-intensive than sediment sampling. Therefore, water profiling and sampling were performed at a representative subset of stations. The six CTD–rosette stations were selected to maximize interpretive value by capturing the principal cross-shelf gradients (nearshore to ~30 m), the dominant hydrographic regimes, and the key sedimentary facies identified during prior acoustic surveys, while remaining consistent with ship-time constraints for vertical profiling and multi-parameter laboratory determinations. This design provides a robust characterization of the vertical biogeochemical structure of the water column alongside a spatially resolved sediment baseline; stations without water sampling are interpreted within the same hydrographic context constrained by the targeted CTD profiles.

2.3.1. Nutrients (PO₄³⁻, SiO₄⁴⁻, NO₂⁻, NO₃⁻, NH₄⁺)

From each rosette cast, aliquots were taken for nutrients (approx. 0.5 L per depth). In total, 18 water sub-samples were preserved by freezing at −24 °C and subsequently analyzed at NIRD GeoEcoMar (Constanta) following standard seawater methods [36].

• Phosphate (PO₄³⁻): Formation of phosphomolybdate in an acid medium and reduction with ascorbic acid to a blue complex; absorbance at 880 nm (PerkinElmer Lambda 35 UV-VIS);
• Silicate (SiO₄⁴⁻): Silicomolybdate method with ascorbic acid reduction; absorbance at 810 nm for low to moderate concentrations or 660 nm for high concentrations;
• Nitrite (NO₂⁻): Diazotization with sulfanilamide and coupling with N-(1-naphthyl) ethylenediamine; absorbance at 540 nm; linear range 0–10 µmol L⁻¹;
• Nitrate (NO₃⁻): Homogeneous reduction to nitrite with hydrazine sulfate in the presence of Cu²⁺, followed by the nitrite procedure above (measurement at 540 nm). Results represent (NO₃⁻ + NO₂⁻); nitrite was measured separately and subtracted;
• Ammonium (NH₄⁺): Indophenol blue formation from monochloramine in moderately alkaline medium (pH 8–11.5) with phenol and nitroprusside; absorbance at 630 nm.

2.3.2. Chlorophyll-a

Between 1 and 5 L of seawater per depth were filtered shipboard through 47 mm nitrocellulose Millipore filters (0.8 µm). Filters were frozen at −24 °C and analyzed by UV-VIS spectrophotometry (PerkinElmer Lambda 35, Markham, ON, Canada) after extraction in 90% acetone, following Aminot et al. [37]. Chlorophyll-a concentrations were calculated using the trichromatic method and the Jeffrey & Humphrey [38] equations, based on absorbances at 750, 664, 647, and 630 nm (optical pathlength 5 cm).

2.3.3. Total Organic Carbon (TOC) and Total Nitrogen (TN)

Water for TOC/TN (approx. 50 mL per depth; 18 sub-samples) was collected in pre-combusted glass vials, left unfiltered, acidified with HCl to remove inorganic carbon, stored at 4 °C, and analyzed within 7 days. TOC was determined by high-temperature catalytic oxidation (HTCO) with NDIR detection as TC–TIC after acidification and CO₂ purge. Thus, the values represent bulk TOC (DOC + POC). TN was measured on the same unfiltered aliquots by HTCO with chemiluminescent/NDIR detection (instrument default) and is reported as bulk TN (dissolved + particulate nitrogen). Certified standards, blanks, and replicates were run for each batch; instrument calibration and MDLs followed manufacturer recommendations. To minimize particle settling, vials were gently inverted immediately prior to injection.

2.3.4. Dissolved Mercury (Hg)

For total dissolved Hg, six refrigerated (4 °C) samples were analyzed with a DMA-80 Milestone direct mercury analyzer, which thermally decomposes the sample, catalytically reduces Hg, and quantifies it by atomic absorption spectrophotometry. The method detection limit was 0.003 ng Hg, with high reproducibility. Quality control employed the certified reference material CRM 3163.

2.3.5. Dissolved Metals by ICP-MS

Six seawater samples preserved with HNO₃ (10%, v/v) and stored at 4 °C were analyzed for Cr, Mn, Co, Ni, Cu, Zn, As, Cd, Pb, Fe, and Hg by inductively coupled plasma mass spectrometry (ICP-MS). An argon plasma provided ionization prior to mass spectrometric separation and quantification.

2.3.6. Dissolved PAHs by UPLC-FLD

PAHs in seawater (six samples) were quantified using an ACQUITY UPLC system with fluorescence detection (Waters). Separation employed a Waters PAH 4.6 × 250 mm column at 30 °C under gradient elution (acetonitrile/water, 1.2 mL min⁻¹). Excitation/emission wavelengths were programmed for each target PAH, as optimized in Chizhova et al. [39]. Identification was based on retention times and fluorescence spectra; quantification used external calibration.

2.4. Surface Sediments: Collection, Preservation, and Analyses

Surface sediments were collected at 25 stations using a Van Veen grab (sampled area 0.135 m²) (Figure 2). Sub-samples were transferred to plastic containers and stored at room temperature until analysis in the GeoEcoMar geochemistry laboratory [40]. For PAHs, the upper 0–5 cm layer of sediment was separately sampled at six stations using the same grab and transferred into 0.5 L amber glass jars and immediately frozen at −20 °C. Prior to organic contaminant analysis, sediment was homogenized, air-dried at 25 °C for aprox. 1 week, and sieved at 150 µm. Moisture content was determined by oven drying at 105 °C for 4 h. When the residual moisture was <1%, no correction was applied to the PAH concentrations.

2.4.1. Carbonates and Total Organic Carbon (TOC)

• Calcium carbonate (CaCO₃): Determined volumetrically from CO₂ released upon acidification with 0.5 N HCl and back-titration of excess acid with 0.5 N NaOH using phenolphthalein, after Black [41];
• Sedimentary oxidizable organic carbon (modified Walkley–Black). A known mass of dried, sieved sediment was subjected to wet oxidation with excess K₂Cr₂O₇ in concentrated H₂SO₄ (with Ag₂SO₄ to suppress chloride interference). The exothermic heat of acid dilution (with gentle external heating as needed) sustained the reaction for 30 min. Residual dichromate was back-titrated with 0.5 N Fe(NH₄)₂(SO₄)₂ (Mohr’s salt) using an o-phenanthroline/ferroin or diphenylamine(/-sulfonate) redox indicator. Reagent blanks and certified reference material (e.g., SRM 2702) were processed in parallel; organic carbon was calculated from dichromate consumption and corrected using a laboratory recovery factor derived from CRMs to account for incomplete oxidation of refractory carbon in the classical procedure. Results are reported as TOC (%C). Where organic matter (OM) was needed, OM was derived from OC using an accepted conversion factor (e.g., OM = 1.724 × OC). Highly calcareous samples were pre-treated with dilute HCl to remove carbonates prior to oxidation.

2.4.2. Major, Minor, and Trace Elements, Total Hg

Elemental concentrations (Fe, Mn, Cr, V, Ni, Cu, Zn, As, Rb, Zr, Pb, and others) were determined by energy-dispersive X-ray fluorescence (EDXRF) using a SPECTRO XEPOS C benchtop spectrometer. Qualitative identification relied on characteristic fluorescent X-ray energies; quantitative results were obtained from emission intensities using instrument calibrations. Total Hg in the sediments was measured with a DMA-80 Milestone analyzer by thermal decomposition and AAS detection (LOD 0.003 ng Hg). Quality control used SRM 2702 (Inorganics in Marine Sediment).

2.4.3. Sedimentary PAHs (QuEChERS-Type Extraction; UPLC-FLD)

PAHs in surface sediments (six stations) were analyzed following [42]. One gram of dried, sieved sediment was weighed into a 15 mL centrifuge tube, then extracted with 1 mL ultrapure water and 4 mL acetone. After vortexing at 3000 rpm for 30 s, tubes were sonicated for 15 min and centrifuged at 4000 rpm for 5 min. A 1 mL aliquot of the supernatant was transferred to a 2 mL tube containing 25 mg PSA and 90 mg MgSO₄ (dSPE cleanup), vortexed (3000 rpm, 30 s), and centrifuged (12,000 rpm, 5 min). The final supernatant was membrane-filtered (Milipore, Burlington, MA, USA) and analyzed by UPLC-FLD, as in Section 2.3.6 (column, temperature, gradient, and wavelength programming identical). Identification used target retention times and fluorescence spectra; quantification used external standards.

2.4.4. Gamma Ray Spectrometry of Sediments

Gamma ray spectrometry was performed in order to evaluate natural series concentrations and test the anthropogenic components, but only on sediment samples, since the low radionuclide concentrations in seawater were not compatible with the relatively low detection efficiency of the HPGe detector and associated setup. Although the HPGe system offers high resolution, gamma detection limits are orders of magnitude higher than those achievable by ICP-MS under standard conditions, even for advanced methodologies [43]. The spectrometric system used consisted of a liquid nitrogen-cooled extended range ORTEC detector of 40% relative efficiency and 2.1 keV FWHM at 1332 keV, with a DSPEC integrated signal analysis module, with the low background ensured by a 10 cm thick Pb shield and an inner 4 mm Cu layer to reduce characteristic X-ray contribution. Typical spectrum acquisition times were 12 h, with a 1 Bq/kg or lower detection limit for the nuclides of interest. The first measurement sequence was performed for the natural spectrum, i.e., samples were dried (under 1% relative humidity), milled for uniformity, as the latter is very important [44,45], and sealed in identical, standardized containers. The second sequence collected the spectra post neutron irradiation in order to check for which elements and isotopes this process would reveal significant changes for (semi)quantitative determinations. Although awaiting investigations with modern systems [46], the old neutron source used is not yet characterized. Still, it has a “twin” described in [47]. However, as the manufacturing process goes back at least to 1976, the uncertainties are high, and all we can fairly state is that the neutron flux is of some 1 to 5 × 10⁵ n/s in 4π with a mean energy of 3 Mev, from some 9 MeV to thermal neutrons. Most of the neutrons are thermalized within some 10 cm of surrounding water, so we placed the volume samples of 10 cm in diameter in the tank right near the source to obtain the entire spectrum, from high energy to thermal. This way, it is possible to test several reactions in a single iteration, for the cross sections vary by several orders of magnitude with neutron energy for the range of isotopes present in such samples. However, the determinations are just semiquantitative. Spectrum analysis was carried out with Interspec [48] and GaDeTool [49]. Self-attenuation corrections and coincidence-summing corrections [50] for efficiency estimation were performed with GESOPECOR 5.0. [51]. Special attention was paid to the ²¹⁰Pb peaks [52].

Although radioactivity-based studies in this field are not new [53,54,55], gamma ray analysis opens the way to radiological investigations from the natural series and non-chain elements to neutron-induced supplementary characterization [56]. Some of those studies are part of a larger effort to study the Danube from the Delta–River–Sea system perspective [57]. In situ measurements will soon follow to help correlate the data [58] and enable covariance studies [59].

In order to evaluate neutron activation effects, we first selected the most abundant isotopes for the elements listed in the XRF table. Then, a selection of the most probable reactions which lead to a significant gamma ray yield (or at least an a priori interference-free part of the spectrum) and lifetimes of at least a few minutes (so they can be recorded in the spectrum) was made for metals and metalloids from Ca to Pb. Some nuclei are of particular interest for current research topics, such as Si [60,61], for their structure and gamma emissions, or Os [62], for the study of their concentrations in seawater [63]. Among the pollutants, Cd isotopes exhibit a transition from U(5) to O(6) symmetry, which has an extensive region near A = 130 [64]—as in the case of La [64], another oligoelement of which ecotoxicity is of current interest [65]. Bismuth [66] is another current nuclear reaction research topic. Following neutron activation of ²⁰⁸Pb, the ²⁰⁹Pb decays shortly to ²⁰⁹Bi, which may capture a neutron to form ²¹⁰Bi in an excited state. Those more complex processes are meant to test the detection limits. In short, nuclide identification is far from being straightforward, as there are many reactions that lead to energies very close to each other.

The energies we sought for, namely associated to the most probable reactions, were as follows:

-

⁵²Cr(n, 2n)⁵¹Cr: 320.08 keV;

-

⁴⁰Ca(n, p)⁴⁰K: 1460.82 keV;

-

²⁷Al(n, γ)²⁸Al: 1778.99 keV;

-

²⁸Si(n, p)²⁸Al: 1778.99 keV;

-

⁴⁸Ti(n, n′γ)⁴⁸Ti: 1037.54 keV; ⁴⁸Ti(n, p)⁴⁸Sc: 983.53 & 1312.12 keV;

-

⁵¹V(n, γ)⁵²V: 1434.06 keV; ⁵¹V(n, p)⁵¹Ti: 320.08 keV; ⁵¹V(n, α)⁴⁸Sc: 983.53 & 1312.12 keV;

-

⁵⁵Mn(n, γ)⁵⁶Mn: 375.18, 846.76 keV;

-

⁵⁶Fe(n, n′γ)⁵⁶Fe 846.76 keV; ⁵⁶Fe(n, γ)⁵⁷Fe: 692.03 keV; ⁵⁶Fe(n, p)⁵⁶Mn: 375.18 keV;

-

⁵⁸Ni(n, p)⁵⁸Co: 810.76 keV; ⁶⁰Ni(n, γ)⁶¹Ni: 282.96 & 656.01 keV;

-

⁶³Cu(n, γ)⁶⁴Cu: 1345.77 keV;

-

⁶⁴Zn(n, 2n)⁶³Zn: 669.64 & 962.07 keV; ⁶⁴Zn(n, α)⁶¹Ni: 282.96 & 656.01 keV;

-

⁷⁵As(n, γ)⁷⁶As: 559.09 keV;

-

⁸⁸Sr(n, n′γ)⁸⁸Sr 1836.06 keV;

-

²⁰²Hg(n, γ)²⁰³Hg: 341.5 & 591.4 keV;

-

²⁰⁸Pb(n, p)²⁰⁸Tl: 583.19 & 2614.51 keV;

-

¹¹³Cd(n, γ)¹¹⁴Cd: 558.46 & 805.89 keV;

-

¹³⁹La(n, γ)¹⁴⁰La: 1596.21 keV;

-

¹⁹⁰Os(n, γ)¹⁹¹Os: 129.43 keV;

-

²⁰⁹Bi(n, γ)²¹⁰Bi: 265.60 keV.

All nuclear data was retrieved from the Brookhaven National Laboratory repositories [67]. In order to select the plausible outcomes from the list, we cross-checked with the XRF data again.

2.5. Quality Assurance and Quality Control (QA/QC)

All glassware and sampling equipment were acid-cleaned prior to use. Field and laboratory blanks accompanied each analytical batch where applicable. Analytical performance was verified using certified reference materials: CRM 3163 for dissolved mercury and SRM 2702 for sediment inorganic constituents, together with calibration standards and replicate analyses. Instrumental detection limits and linear ranges followed manufacturer recommendations and cited methods; sample handling and storage conformed to best practice to minimize contamination and analyte loss.

2.6. Statistical Analysis

All statistical analyses were performed using PRIMER 7 with the PERMANOVA+, add-on software package [68], Ocean Data View (ODV) version 5.8.3 [69], and Past 4 software [70]. Prior to PCA/FA, the variables were z-standardized; communality > 0.6 and loadings |≥0.6| were interpreted as strong. To test H3 against lithogenic confounding, we inspected partial correlations and Fe-normalized residuals (when applicable) for Pb, Hg, and ΣHMW-PAH. Station scores on PC1–PC2 were mapped to evaluate hotspot persistence independent of matrix effects.

3. Results

3.1. Water Analysis

3.1.1. Nutrients, Chlorophyll-a, Total Organic Carbon (TOC), and Total Nitrogen (TN)

Nutrient levels were within the expected range for the warm season, with spatial variability driven by hydro-meteorological conditions and in-column biogeochemical processing (

Table 1. Concentration of nutrients chlorophyll-a, TOC, and total nitrogen in the water samples studied.

Measuring Station	Depth m	PO4 [μMol]	SiO4 [μMol]	NO2 [μMol]	NO3 [μMol]	NH4 [μMol]	Chl a [μg/L]	TOC [mg/L]	TN [mg/L]
S01	0	0.19	5.98	0.09	0.03	2.97	0.623	4.85	0.43
	16	0.07	2.51	0.07	0.44	1.20	1.164	3.35	0.26
	25	0.31	11.23	0.06	0.46	9.51	0.554	3.49	0.26
S07	0	0.02	24.53	0.03	0.27	8.56	0.663	3.03	0.24
	15	0.10	6.49	0.06	0.39	1.93	1.196	3.32	0.32
	18	0.04	6.90	0.01	0.39	0.89	0.628	3.09	0.26
S06	0	0.02	8.81	0.05	0.47	0.41	0.816	2.83	0.25
	16	0.09	6.54	0.08	0.47	0.56	1.035	2.2	0.22
	23	0.45	13.42	0.09	0.82	0.66	0.748	7.35	0.32
S13	0	0.03	5.57	0.05	0.34	1.30	0.982	1.79	0.16
	15	0.52	14.95	0.18	1.08	0.50	1.283	2.59	0.32
	25	0.59	15.88	0.29	0.29	1.35	0.841	1.82	0.18
S20	0	0.18	3.45	0.02	0.33	0.78	0.512	2.82	0.29
	12	0.11	4.29	0.03	0.63	1.68	1.563	2.52	0.31
	30	0.24	7.13	0.07	0.56	2.94	0.909	2	0.24
S23	0	0.24	5.64	0.03	0.34	2.52	0.447	4.09	0.25
	16	0.18	2.91	0.05	0.64	19.33	1.336	1.65	0.22
	30	0.19	5.75	0.11	0.57	7.83	1.009	1.45	0.25

). Concentrations spanned 0.02–0.59 µM (PO₄³⁻), 2.52–24.53 µM (SiO₄⁴⁻), 0.03–1.08 µM (NO₃⁻), 0.01–0.29 µM (NO₂⁻), and 0.42–19.33 µM (NH₄⁺).

Vertically, phosphate and silicate maxima generally occurred in bottom layers at most stations, consistent with remineralization and/or benthic fluxes. The main exception was station S07, where the silicate maximum was found at the surface, suggesting recent surface inputs or mixing. Nitrite and nitrate peaked at S13 in deeper strata (15–25 m), while the highest ammonium concentration occurred at S23 (16 m). Across most stations, ammonium was the dominant inorganic nitrogen form, indicating active organic matter mineralization and a preferential phytoplankton uptake of oxidized nitrogen species.

Chlorophyll-a ranged from 0.447 to 1.563 µg L⁻¹ (Table 1), with a pronounced subsurface maximum at S20 (12 m). Concentrations exceeding 1 µg L⁻¹ were observed consistently within the chlorophyll maximum layer (12–16 m) across stations, pointing to a typical summer deep-chlorophyll-maximum structure.

TOC displayed a broadly homogeneous vertical distribution, with two notable departures: a high value at S06 (7.35 mg L⁻¹ at 25 m) and an elevated surface value at S23 (4.09 mg L⁻¹). Surface TOC tended to be higher in the northern sector, consistent with an allochthonous organic matter contribution plausibly linked to Danube influence rather than enhanced local primary production. TN varied narrowly between 0.16 and 0.43 mg L⁻¹, with slightly higher values at depth (e.g., S06, 23 m: 0.32 mg L⁻¹), a pattern compatible with nitrogen limitation in the surface waters during summer stratification.

The vertical nutrient structure (bottom-layer PO₄³⁻, SiO₄⁴⁻, and NH₄⁺ maxima; subsurface chlorophyll peak) and TOC/TN patterns are characteristic of seasonally stratified coastal waters: remineralization and benthic release at depth, efficient surface uptake of oxidized nitrogen, and a deep-chlorophyll maximum. Spatial TOC anomalies and northern surface enrichment point to riverine/terrestrial inputs superimposed on in situ processes (Figure 3 and Figure 4).

3.1.2. Dissolved Metals in Surface Waters

Essential elements showed the following ranges in surface waters: Fe = 1.12–5.91 ppb, Zn = 4.25–10.77 ppb, and Cu = <0.45–38.37 ppb (

Table 2. Concentrations of inorganic contaminants in the surface water layer in the studied area.

Station	Cr52, ppb	Mn55, ppb	Co59, ppb	Ni60, ppb	Cu63, ppb	Zn66, ppb	As, ppb	Cd117, ppb	Pb207, ppb	Hg, ppb	Fe56, ppb
SO01	1.81	10.56	0.39	0.99	<0.45	10.77	3.85	<0.54	27.41	0.042	1.53
SO06	3.94	8.82	1.64	0.96	38.37	4.26	0.65	<0.54	12.52	0.045	5.91
SO07	3.52	4.10	1.48	<0.85	<0.45	4.82	1.27	1.24	8.43	0.038	2.73
S13	3.10	5.98	0.23	<0.85	<0.45	<1.07	1.51	1.28	5.49	0.029	2.51
S20	4.04	4.09	0.22	<0.85	<0.45	<1.07	0.54	1.23	4.38	0.032	1.12
S25	3.36	4.11	0.57	<0.85	<0.45	8.94	1.23	1.32	1.80	0.037	3.58

). Maximum Fe and Cu occurred at S06, whereas Zn peaked at S01. Manganese was comparatively homogeneous (4.09–10.56 ppb) with the highest value at the shallow station S01. Among potentially toxic metals, elevated concentrations were predominantly associated with shallow stations: Pb up to 27.4 ppb (S01), Cr up to 4.04 ppb (S20), As up to 3.85 ppb (S01), Cd up to 1.32 ppb (S25), and Co up to 1.63 ppb (S06). Mercury was detected at all stations but remained low (0.03–0.05 ppb). The enrichment of Pb, Cr, As, Cd, and Co at shallow sites, together with high Cu and Zn at S06/S01, indicates nearshore sources (e.g., harbor activities, urban/industrial runoff, and/or resuspension). The uniformly low dissolved Hg suggests limited recent inputs or efficient scavenging from the dissolved phase under prevailing conditions.

3.1.3. Dissolved Organic Contaminants (PAHs) in Surface Waters

The sum of 16 target PAHs in surface water ranged from 134 to 347 ng L⁻¹, with the highest totals at S01, S06, and S07 (

Table 3. Concentrations of organic contaminants in the surface water layer of the investigated area (ng/L), ND: not detected.

Nr.crt.	Component (ng/L)	Station
		S01	S06	S07	S13	S20	S25
1	Naphthalene	109.20	107.60	109.20	56.20	16.20	28.60
2	Acenaphthene	63.60	70.40	68.80	38.00	11.00	19.60
3	Fluorene	14.80	25.00	24.80	17.40	8.20	11.60
4	Phenanthrene	76.40	87.20	85.00	96.60	64.20	86.40
5	Anthracene	5.00	6.60	6.80	6.00	3.60	5.00
6	Fluoranthene	12.00	15.40	12.60	18.80	13.20	18.40
7	Pyrene	16.40	15.20	9.60	17.80	12.00	11.20
8	Benzo (a) anthracene	0.20	2.40	0.20	2.80	3.20	0.60
9	Chrysene	1.80	0.80	0.60	1.40	1.00	1.40
10	Benzo (b) fluoranthene	0.60	0.40	1.40	0.60	0.40	0.60
11	Benzo (k) fluoranthene	ND	1.40	1.00	0.20	0.20	0.20
12	Benzo (a) pyrene	10.60	7.00	7.80	8.20	0.80	1.60
13	Dibenzo(a,h)anthracene	7.80	7.60	5.60	4.20	ND	29.00
14	Benzo(g,h,l)perylene	25.80	ND	ND	ND	ND	ND
15	Indenol(1,2,3-cd) pyrene	ND	ND	ND	0.00	ND	0.00
	Sum of PAH components (ng/L)	344.20	347.00	333.40	268.20	134.00	214.20

), likely reflecting maritime/harbor influences. The PAH mixture was dominated by low-molecular-weight (LMW, 2–3 rings) species—naphthalene (16.2–109.2 ng L⁻¹), acenaphthene (11–70.4 ng L⁻¹), and fluorene (8.2–25.0 ng L⁻¹)—which are more soluble and typically indicative of petrogenic inputs (e.g., fuel/oil handling, and shipping). High-molecular-weight (HMW) PAHs (e.g., fluoranthene, benzo[a]anthracene, benzo[a]pyrene, dibenzo[a,h]anthracene, and indenol [1,2,3-cd] pyrene) were generally one order of magnitude lower (<1–18.8 ng L⁻¹) or not detected. Two noteworthy exceptions suggest localized pyrogenic signatures: benzo(g,h,i) perylene at S01 (25.8 ng L⁻¹) and dibenzo[a,h]anthracene at S25 (29.0 ng L⁻¹). Basin-wide, LMW dominance supports a largely petrogenic signal associated with maritime traffic and port operations. The isolated HMW peaks at S01 and S25 point to episodic combustion-related inputs (pyrogenic) superimposed on the broader petrogenic background.

3.2. Sediment Characterization (Surface Layer)

Sediments from the investigated sector exhibit pronounced lithological and textural heterogeneity, reflecting the combined influence of regional coastal geology, Danube-derived detrital supply, and alongshore transport. The adjacent onshore and nearshore basement is dominated by Dobrogea lithologies (metamorphic and sedimentary successions, locally including carbonate units), with additional contributions from Quaternary loessic and coastal deposits. Weathering and erosion of these rock types, together with riverine inputs, provide the siliciclastic fraction that dominates the shelf sediments. This geological context supports the interpretation that the trace-element inventory is largely governed by terrigenous carriers (aluminosilicates, Fe-Mn oxyhydroxides, and heavy-mineral phases), with a local superposition of anthropogenic inputs near harbors and urbanized coastlines. At the station scale, facies range from micaceous sands and sandy muds to fine-grained muddy deposits, with local biogenic accumulations dominated by molluscan shell debris. The frequent occurrence of mica points to a persistent continental detrital flux transported by fluvial discharge and coastal drift. Many cores display a bipartite vertical structure: an oxidized, semi-fluid sandy-mud surface veneer overlying more compact grey-to-black, organic-enriched muds locally overprinted by bioturbation. This stratification is consistent with an oxic-to-suboxic redox gradient within the upper sediment column, primarily controlled by oxygen penetration depth and the balance between organic matter supply and remineralization. Facies distribution and benthic features indicate systematic bathymetric and hydrodynamic zonation across the study area. The inner shelf is dominated by sands and sandy muds deposited under moderate hydrodynamic energy, where episodic resuspension and benthic reworking limit the preservation of fine particles and associated organic matter. Basinward, sediments grade into a lower-energy sublittoral setting characterized by finer-grained, more weakly oxygenated muds, slower sedimentation dynamics, and comparatively enhanced preservation of organic material. Locally developed biogenic facies, composed largely of Mytilus bioclasts, record biologically mediated sediment accumulation associated with dense benthic communities and/or episodic mortality and redeposition events.

Overall, the sedimentary architecture is consistent with a low-to-moderate energy shelf system in which alongshore transport, intermittent resuspension, and gradual deposition interact to produce strong spatial coupling among grain size, redox conditions, and benthic activity. This coupling is central for interpreting the observed terrigenous control on trace-element distributions and provides a robust framework for evaluating present-day benthic habitat quality and recent environmental change on the Romanian Black Sea shelf.

Alongshore circulation on the Romanian Black Sea coast is strongly influenced by the Danube buoyant discharge, which generates a persistent longshore current largely confined to the surface layer, even under relatively low river flow. Superimposed wind forcing, most frequently from NE and SE, modulates current intensity and vertical penetration; winds blowing parallel to the coastline can produce the strongest currents and drive mixing/resuspension down to roughly 20–30 m of water depth, thereby promoting sediment redistribution and alongshore transport [71].

3.2.1. Carbonates and Total Organic Carbon (TOC)

Calcium carbonate (CaCO₃). CaCO₃ ranged from 7.6 to 29.9% (mean 13.2%), with higher values at nearshore stations (e.g., S01 = 28.74%, S05 = 28.04%, S06 = 29.87%) and a clear decrease offshore (e.g., S18 = 10.09%, S19 = 7.60%, S20 = 7.96%) (Figure 5;

Table 4. Concentrations of major and minor components and trace elements from the superficial sediments within the studied perimeter (MN 270).

Station	Depth [m]	TOC, %	CaCO3, %	Ti, ppm	V, ppm	Cr, ppm	Mn, %	Fe, %	Ni, ppm	Cu, ppm	Zn, ppm	As, ppm	Hg, ppm	Pb, ppm
S01	27	0.552	28.74	2790	57.9	75.6	0.05	2.28	33.59	18.45	55.7	7.89	0.06	17.98
S02	20.5	0.219	19.25	3223	46.82	86.1	0.05	2.18	30.35	9.14	41.77	6.48	0.02	13.82
S03	15	0.109	9.81	4884	45.51	91.3	0.06	2.44	26.65	7.35	43.63	4.88	0.02	13.09
S04	17.5	0.163	14.11	3515	52.1	90.8	0.05	2.29	31.61	11.41	47.17	5.98	0.02	14.75
S05	22	0.428	28.04	2555	52.3	76.2	0.06	2.3	30.57	15.05	48.81	10.9	0.04	14.95
S06	26.5	0.485	29.87	2462	50.6	63.3	0.04	2.04	29.01	15.79	51	5.62	0.06	15.82
S07	20.5	0.287	9.28	3525	49.92	89.5	0.05	2.28	31.45	13.15	51.5	5.36	0.04	16.33
S08	22	0.391	10.87	4411	62.1	121	0.06	2.77	40.94	17.14	64.7	6.91	0.05	18.36
S09	24.5	0.359	9.51	4427	57.4	102	0.06	2.54	35.81	17.8	60.1	7.12	0.06	19.08
S10	27.5	0.779	10.15	3569	64.7	101	0.05	2.67	39.89	24.5	69.6	7.14	0.08	21.32
S11	26	0.610	8.99	3617	69.2	97.7	0.05	2.83	43.06	25.56	73	7.94	0.08	22.51
S12	23.4	0.785	10.94	3671	74.9	99.1	0.05	3.03	45.29	28.01	78.6	8.69	0.08	24.2
S13	20.8	0.342	9.35	3479	70.5	89.7	0.05	2.88	43.4	24.14	74.3	8.24	0.07	23.31
S14	30	0.249	10.37	4887	58	90.3	0.06	2.65	36.33	15.45	59.5	6.42	0.04	17.65
S15	25	0.367	8.00	4813	57.4	102	0.05	2.47	33.78	13.91	54.1	6.29	0.04	16.23
S16	23	0.371	10.71	4303	61.3	116	0.06	2.63	37.43	19.62	64	7.5	0.06	19.16
S17	22.4	0.525	9.35	3731	79.4	96	0.05	3.3	52.1	38.53	89	8.65	0.09	25.47
S18	25.7	0.501	10.09	4236	69.8	91.5	0.05	2.8	42.9	23.72	71.8	6.42	0.07	21.77
S19	30.2	0.484	7.60	3837	72.2	90	0.05	2.83	44.45	23.03	72.8	6.64	0.07	22.55
S20	33	0.302	7.96	3353	65.4	87.3	0.04	2.6	40.1	22.5	68.8	6.5	0.08	21.3
S21	33.3	0.167	22.43	2974	61.7	85.4	0.05	2.41	36.74	19.62	61.1	8.31	0.06	19.96
S22	26	0.416	11.35	3728	68.3	96.6	0.05	2.85	44.32	21.01	65.8	8.34	0.05	18.75
S23	34.6	0.637	11.01	3583	81.4	99.9	0.05	3.28	51.1	30.04	86.4	8.81	0.09	25.66
S24	38	0.964	12.97	3446	67.9	93.7	0.05	2.77	43.7	24.28	73.6	7.79	0.07	22.54
S25	36	0.675	8.89	4042	76.1	120	0.06	3.25	47.52	23.36	77.8	8.32	0.06	24.22

). Most nearshore sites fall within the non-carbonate terrigenous-to-slightly calcareous terrigenous classes (<10–30% CaCO₃), Emelyanov & Shimkus [72], with the lowest value at S19. The cross-shore pattern indicates bioclastic enrichment in the surf/shoreface, where shell debris accumulates, and dilution by siliciclastic sands seaward.

TOC. TOC varied from 0.11 to 0.96% (mean 0.45%) and was generally ≤1% across stations (Table 4). Spatially, TOC shows a patchy distribution with no monotonic cross-shore trend: elevated values occur both nearshore (e.g., S10 = 0.779%, S12 = 0.785%) and offshore (e.g., S24 = 0.964%). This pattern is consistent with textural control (finer fractions favoring organic matter preservation), intermittent resuspension/settling, and local hydrodynamics.

Overall, the nearshore CaCO₃ maximum, together with low, sand-dominated TOC, reflects a shoreface bioclastic enrichment and inner-shelf siliciclastic dilution. An overview of CaCO₃ (%) and TOC (%) in surface sediments is provided in Figure 5.

3.2.2. Major and Trace Elements

Iron (Fe) and matrix indicators. Fe shows relatively low spatial variability (2.04–3.30%) compared with CaCO₃ (Table 4). The correlation matrix indicates strong positive associations of Fe with V (r = 0.927), Ni (0.945), Cu (0.819), Zn (0.921), and Pb (0.872), and a moderate one with Cr (0.598), supporting a predominantly terrigenous control (fine aluminosilicates/Fe-oxides) on these metals (

Table 5. Correlation matrix.

Variables	TOC	CaCO3	Mg	Al	Ti	V	Cr	Mn	Fe	Ni	Cu	Zn	As	Sr	Hg
TOC	1
CaCO3	−0.066	1
Mg	0.127	−0.801	1
Al	0.354	−0.785	0.919	1
Ti	−0.188	−0.739	0.657	0.479	1
V	0.618	−0.441	0.491	0.752	0.060	1
Cr	0.197	−0.686	0.611	0.617	0.651	0.414	1
Mn	−0.237	−0.236	0.289	0.207	0.571	0.072	0.530	1
Fe	0.516	−0.608	0.615	0.812	0.318	0.927	0.598	0.301	1
Ni	0.593	−0.497	0.551	0.801	0.105	0.982	0.474	0.091	0.945	1
Cu	0.650	−0.300	0.456	0.709	−0.106	0.925	0.239	−0.069	0.819	0.911	1
Zn	0.642	−0.472	0.528	0.789	0.079	0.979	0.419	0.039	0.921	0.972	0.955	1
As	0.432	0.226	−0.012	0.154	−0.381	0.522	0.088	0.272	0.447	0.492	0.545	0.472	1
Sr	−0.228	0.938	−0.875	−0.885	−0.651	−0.564	−0.693	−0.206	−0.697	−0.619	−0.462	−0.611	0.026	1
Hg	0.674	−0.228	0.324	0.595	−0.185	0.836	0.170	−0.199	0.658	0.796	0.918	0.890	0.438	−0.400	1
Pb	0.652	−0.428	0.448	0.728	−0.001	0.965	0.363	−0.028	0.872	0.942	0.939	0.982	0.475	−0.568	0.910

).

Chromium (Cr). Cr spans 63.3–121 ppm; the highest values occur at the shallow Eforie stations (S08 = 121 ppm; S10 = 101 ppm; S15 = 102 ppm; S16 = 116 ppm; and S25 = 120 ppm, Table 4). Significant positive correlations for Ti-Cr (r = 0.651) and Fe-Cr (0.598) further indicate a lithogenic (terrigenous) origin, with Cr hosted by heavy-mineral/oxide phases in the siliciclastic fraction.

Essential metals (Ni, Cu, Zn). The ranges are Ni 26.7–52.1 ppm, Cu 7.35–38.53 ppm, Zn 41.8–89.0 ppm (Table 4). Spatially, these elements tend to be higher near the coast (Eforie) and at a few deeper stations from the same sector, with secondary increases toward the open sea near Costinești. Strong correlations with Fe (Ni 0.945; Cu 0.819; Zn 0.921) suggest a common, mostly natural association with Fe-oxide/hydroxide coatings and fine terrigenous particles. Local anthropogenic influences nearshore may superimpose on this natural background.

Potentially toxic metals (Pb, As, Hg).

• Pb: 13.1–25.66 ppm, maximum at S23 (nearshore);
• As: 4.88–10.9 ppm, maximum at S05 (closest to shore);
• Hg: 0.016–0.090 ppm, maximum at S23 (nearshore).

Positive Pb-Hg-Fe family correlations (e.g., Hg-Pb r = 0.910, Hg-Zn 0.890, and Pb-Zn 0.982) indicate co-enrichment in the fine terrigenous/oxide fraction, while the nearshore maxima at S23 point to anthropogenic inputs (harbor/urban) superposed on natural carriers.

Arithmetic means for As, Cr, Cu, Pb, Hg, and Zn remain below Romanian sediment-quality criteria [20]. Ni averages 38.88 ppm, slightly exceeding the 35 ppm threshold. Given the high Ni-Fe (0.945) and Ni-V (0.982) correlations and the regional geology (green schists, basic rocks), the exceedance is plausibly natural, via adsorption/coprecipitation with Mn/Fe oxides and lithogenic enrichment rather than point-source pollution.

The metal suite is lithogenically dominated (Fe-oxide-associated), with nearshore hotspots (S23; Eforie) signaling local anthropogenic contributions for Pb and Hg. Elevated Cr in shallow Eforie stations is consistent with terrigenous inputs rich in oxide and heavy-mineral phases.

Metals such as Ni, Cu, and Zn are essential to various biological processes, but may also become toxic at high concentrations. Analysis of those elements’ spatial distribution indicates a higher accumulation within the coastal zone stations (low depths) of Eforie, but also in the case of a few greater depths from the same area and open sea around Costinești. Statistical correlations between those metals and their association with Fe suggest a possible common, mostly natural, origin (Table 5).

Hg and Pb, which are well-known for their high toxicity even at low concentrations, had their highest values in stations from the coastal zone of Eforie and Costinești, with a maximum of 0.09 ppm Hg and 25.66 ppm Pb for station S23. Positive statistical correlations between those elements suggest they have a common anthropic source within the coastal area.

The highest arsenic (As) concentration was measured at the station closest to the shoreline (S05; 10.9 ppm), showing a nearshore enrichment pattern similar to that observed for Hg and Pb in the surface sediments. On a general basis, high toxicity metals did not exceed the legal limits; their lean values for As, Cr, Cu, Pb, Hg, and Zn were, respectively, 7.33 ppm, 94.08 ppm, 20.10 pm, 19.63 ppm, 0.06 ppm, and 64.18 ppm (Figure 6).

3.2.3. Polycyclic Aromatic Hydrocarbons (PAHs) in Sediments

Σ16 PAHs range 41.91–88.16 µg kg⁻¹ (mean 54.29 µg kg⁻¹), with the maximum at S06 (

Table 6. Polycyclic aromatic hydrocarbon concentrations in surface sediments from the area under investigation (MN 270).

No.	Component (µg/kg)	Station
		S01	S06	S07	S13	S20	S25
1	Naphthalene	17.91	25.65	17.67	11.24	18.29	19.35
2	Acenaphthene	8.36	14.04	8.08	8.69	9.96	8.41
3	Fluorene	1.39	5.55	1.80	1.53	1.52	1.42
4	Phenanthrene	2.19	13.33	4.29	0.20	0.41	0.71
5	Anthracene	0.30	0.61	0.60	0.82	0.71	0.51
6	Fluoranthene	1.69	3.13	3.19	1.53	3.25	1.42
7	Pyrene	2.59	5.66	5.09	0.41	0.41	0.41
8	Benzo (a) anthracene	1.29	1.51	2.69	1.53	3.76	2.43
9	Chrysene	0.70	6.06	1.30	1.43	2.24	1.11
10	Benzo (b) fluoranthene	0.50	0.50	0.50	4.09	5.39	2.84
11	Benzo (k) fluoranthene	ND	0.50	0.20	1.02	1.12	3.34
12	Benzo (a) pyrene	0.99	0.91	0.10	0.10	0.00	0.10
13	Dibenzo (a,h) anthracene	0.00	0.20	1.10	9.10	ND	0.20
14	Benzo (g,h,l) perylene	0.40	2.42	0.80	0.00	0.20	ND
15	Indenol (1,2,3-cd) pyrene	10.54	8.08	8.78	0.20	0.41	0.71
	PAH Sum of components (µg/kg) ppb	48.84	88.16	56.19	41.91	47.66	42.96

). Three stations show broad detection across the PAH suite.

The mixture is generally LMW-skewed (2–3 rings: naphthalene 11.24–25.65 µg kg⁻¹, acenaphthene 8.08–14.04 µg kg⁻¹, and fluorene 1.39–5.55 µg kg⁻¹), indicative of petrogenic inputs (fuel/oil handling, maritime traffic). HMW indicators of pyrogenic origin (combustion) occur at lower levels, but with local spikes:

• Indeno(1,2,3-cd) pyrene = 10.54 µg kg⁻¹ at S01;
• Dibenzo[a,h]anthracene = 9.10 µg kg⁻¹ at S13;
• elevated benzo(b)fluoranthene at S13–S20.

These patterns suggest a mixed signature: a basin-wide petrogenic background with localized pyrogenic inputs (shipping/harbor, combustion aerosols, and resuspension).

Sedimentary PAHs are low to moderate regionally, dominated by petrogenic LMW compounds, consistent with maritime activity. Local HMW enrichments (S01, S13) point to episodic combustion-related deposition nearshore.

A PCA on standardized sediment variables (TOC, CaCO₃, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, As, Hg, and Pb) explained 46.2% (PC1), 17.3% (PC2), and 9.3% (PC3) of the variance (cumulative 72.8%). PC1 grouped Zn-Pb-V-Ni-Cu together with Fe (and Hg), defining a terrigenous/Fe-oxide carrier axis. PC2 contrasted Ti-Cr (heavy-mineral/oxide) against CaCO₃-Mn (carbonate/facies control), while PC3 captured minor residual variability. FA with three factors confirmed this structure: F1 loaded strongly on Zn, Pb, V, Ni, Cu, and Fe (and Hg), F2 opposed Ti-Cr to CaCO₃-Mn, and F3 highlighted a weaker As-dominated component. The eigenvalues and explained variances are listed in

Table 7. Eigenvalues.

Component	Eigenvalue	Variance_%	Cumulative_%
PC1	7.884	58.218	58.218
PC2	2.983	22.025	80.244
PC3	1.304	9.629	89.873
PC4	0.552	4.076	93.949
PC5	0.271	2.005	95.954
PC6	0.203	1.502	97.456
PC7	0.148	1.089	98.545
PC8	0.093	0.687	99.233
PC9	0.054	0.400	99.632
PC10	0.022	0.166	99.798

; the component loadings are presented in

Table 8. Loadings.

	PC1	PC2	PC3	PC4	PC5	PC6	PC7	PC8	PC9	PC10
TOC_%	0.248	0.168	0.106	0.890	0.047	−0.303	0.111	0.006	0.016	−0.027
CaCO3_%	−0.174	0.430	0.344	0.056	0.168	0.142	−0.704	0.302	−0.057	0.037
Ti_ppm	0.031	−0.555	−0.116	0.049	0.463	−0.099	0.003	0.613	−0.165	−0.107
V_ppm	0.357	0.018	−0.016	−0.137	−0.043	−0.169	−0.209	0.101	−0.384	−0.093
Cr_ppm	0.170	−0.444	0.164	0.227	−0.644	0.446	−0.198	0.165	0.057	0.037
Mn_%	−0.030	−0.395	0.619	0.037	0.405	0.135	−0.058	−0.458	0.071	−0.102
Fe_%	0.339	−0.156	0.060	−0.150	0.024	−0.361	−0.204	−0.088	0.172	0.755
Ni_ppm	0.355	−0.023	−0.016	−0.144	−0.143	−0.228	−0.242	−0.001	0.071	−0.401
Cu_ppm	0.344	0.133	−0.038	−0.106	0.176	0.143	0.033	0.137	0.741	−0.271
Zn_ppm	0.361	0.012	−0.056	−0.064	0.097	0.049	−0.109	−0.079	0.020	−0.008
As_ppm	0.189	0.207	0.643	−0.254	−0.156	−0.097	0.503	0.342	−0.140	0.004
Hg_ppm	0.316	0.192	−0.143	0.079	0.289	0.633	0.192	0.090	−0.139	0.330
Pb_ppm	0.355	0.054	−0.068	−0.047	0.071	0.132	−0.046	−0.358	−0.435	−0.230

, and the station scores on PC1-PC2 are in

Table 9. Scores PC1_PC2.

Station	PC1	PC2
S01	−1.781	2.895
S02	−4.540	0.454
S03	−4.644	−2.948
S04	−3.740	−0.399
S05	−2.846	2.392
S06	−3.571	3.986
S07	−2.991	−0.421
S08	0.090	−2.566
S09	−0.678	−1.765
S10	1.556	0.383
S11	2.215	0.338
S12	3.476	0.618
S13	1.818	0.459
S14	−1.494	−2.199
S15	−1.812	−1.892
S16	0.069	−2.002
S17	5.063	0.545
S18	1.411	−0.356
S19	1.715	−0.120
S20	0.649	1.109
S21	−1.035	1.735
S22	0.753	−0.108
S23	4.735	0.608
S24	2.176	1.049
S25	3.405	−1.796

, consistent with the distribution in Figure 7 (biplot PC1-PC2)

A separate PCA including Σ16PAH (6 stations) accounted for 67.3% (PC1) and 15.1% (PC2) of the variance. Σ16PAH co-varied with Fe-oxide-associated metals (e.g., Ni, Zn, Pb, Cu), consistent with petrogenic signatures transported with fine terrigenous carriers, with room for the local pyrogenic pulses indicated by HMW markers. The eigenvalues and percentage variances are reported in

Table 10. SigmaPAH eigenvalues.

Component	Eigenvalue	Variance_%	Cumulative_%
PC1	11.308	67.312	67.312
PC2	2.533	15.079	82.391
PC3	2.161	12.861	95.253
PC4	0.646	3.847	99.100
PC5	0.151	0.900	100.000
PC6	0.000	0.000	100.000

, the variable loadings in

Table 11. SigmaPAH loadings.

	PC1	PC2	PC3	PC4	PC5	PC6
TOC_%	0.090	0.096	0.676	−0.324	−0.435	−0.039
CaCO3_%	−0.221	−0.169	0.513	0.103	0.002	−0.145
Ti_ppm	0.269	0.319	−0.226	−0.121	−0.123	−0.567
V_ppm	0.318	−0.129	0.073	−0.089	0.043	−0.389
Cr_ppm	0.282	0.309	−0.035	−0.235	−0.350	0.124
Mn_%	0.212	0.443	0.276	0.193	0.236	0.487
Fe_%	0.321	0.058	0.015	−0.172	0.169	0.018
Ni_ppm	0.324	−0.054	−0.010	−0.110	0.057	0.209
Cu_ppm	0.284	−0.335	0.034	0.007	0.109	−0.098
Zn_ppm	0.316	−0.139	−0.035	−0.150	0.197	0.052
As_ppm	0.258	−0.117	0.349	0.463	0.273	−0.269
Hg_ppm	0.133	−0.616	−0.072	−0.075	−0.390	0.268
Pb_ppm	0.319	−0.127	−0.016	−0.062	0.193	0.224
SigmaPAH_µg/kg	−0.266	−0.071	0.126	−0.694	0.519	−0.040

, and the station on PAH_Scores PC1_PC2 in

Table 12. SigmaPAH Scores PC1_PC2.

Station	PC1	PC2
S01	−1.366	−0.236
S06	−4.433	−0.864
S07	−2.366	2.529
S13	2.642	−0.941
S20	0.867	−1.738
S25	4.657	1.250

, consistent with the configuration shown in Figure 8 (PCA biplot with Σ16PAH for the six sediment stations). PC1 (terrigenous/metal-rich) explains 58.2% of the variance and shows coherent positive loadings for Fe, V, Ni, Zn, Cu, and Pb (approx. 0.34–0.36), with a weaker but positive loading for Cr (0.17), consistent with a fine terrigenous/Fe-oxide carrier (Table 8). PC2 (Carbonate–siliciclastic contrast) explains 22.0% and is characterized by CaCO₃ loading positively (0.43), opposite to Ti (−0.56), Cr (−0.44), and Mn (−0.40), capturing facies-driven dilution (Table 8). PC3 explains 9.6% and loads mainly on Mn (0.62) and As (0.64), indicating a redox/diagenetic overprint rather than a distinct anthropogenic metal factor (Table 7 and Table 8). In the PAH-augmented PCA (six sediment stations), the metals group is on PC1 (Fe, Ni, Zn, Pb, and V approx. 0.31–0.32), while the nearshore anthropogenic signature is expressed largely on PC2 (e.g., Hg = −0.62, Cu = −0.34) (Table 10, Table 11 and Table 12). Station scores place nearshore sites toward the anthropogenic/nearshore direction and fine-matrix sites along the terrigenous axis (Figure 8; Table 7, Table 8, Table 9, Table 10, Table 11 and Table 12).

3.2.4. Gamma Spectrometric Measurements

Gamma natural spectrum analysis for the investigated samples led to the values from

Table 13. Results of the gamma spectrometric analysis. “SA” stands for specific activity of the nuclide to the left; “Ld” stands for values under the detection limit; “Norm” stands for yields and activities normed according to the general assumptions on the natural series, in the absence of prior information on secular equilibrium.

Sample	Live (s)	Mass (g)	Pb210	SA1_Bq/kg	Th234	SA2_Bq/kg	U5+Ra3	SA3_Bq/kg	Ra6+U5	SA4_Bq/kg	Pb214	SA5_Bq/kg	Bi214	SA6_Bq/kg	Cs137	SA7_Bq/kg	Ac228	SA8_Bq/kg	K40	SA9_Bq/kg	Ti208	SA10_Bq/kg
peak [keV]			46		93		144		186		352		609		662		969		1461		2614
Yield (%)			4.25		4.24		Norm		Norm		35.72		45.44		85.1		15.9		10.66		99.75
efficiency (%)			7.843		9.9103		9.586		8.2826		4.9508		3.1562		2.9624		2.2305		1.6381		0.9878
MN 270 01	36,076	77.88	504	18.61	840	16.78	Ld	Ld	571	4.47	822	14.3	645	7.09	590	8.34	313	24.95	2570	259.57	519	2.64
MN 270 06	33,802	78.37	478	19.12	898	27.48	Ld	Ld	474	3.29	696	12.64	732	8.84	767	11.5	318	27.41	2677	315.8	546	4.91
MN 270 07	36,751	112.3	734	28.89	1034	22.9	185	3.08	798	5.74	1795	22.05	1269	10.44	519	4.99	391	23.11	3733	333.95	732	6.83
MN 270 20	49,770	74.18	973	42.05	1490	40.87	Ld	Ld	890	6.1	1546	21.33	1161	10.18	1304	14.02	526	33.16	5206	528.8	945	9.06
Blank	138,708	n/a	1269	n/a	2531	n/a	318	n/a	1237	n/a	439	n/a	422	n/a	Ld	n/a	217	n/a	4961	n/a	1714	n/a

. The results are within normal limits, with ²⁰⁸Tl activities much lower than those of ²²⁸Ac, which is understandable mostly for surface sediment, given that there is ²²⁰Rn in between them for the ²³²Th chain. ¹³⁷Cs is still noticeable 39 years after the Chernobyl event. ²⁴¹Am and ²³⁹Pu activities were below detection limits. The ²³⁵U&²²⁶Ra convolution at 186 keV assumed the natural activity ratio of 1:21.7 between ²³⁵U and ²³⁸U. The 144 keV convolution assumed secular equilibrium in the ²³⁵U series. The ²³⁵U values found are quite different. Consequently, those hypotheses are not valid, and an extended study would require a more integrated approach of gamma and ICP-MS. With the exception made for those convolutions, the results reported have associated uncertainties of 15% or less, which is good, for the calibration sources alone have a 10% relative uncertainty (Table 13).

Neutron activation results (

Table 14. Results for the neutron-activated emissions, also expressed in Bq/kg. The values are not associated with a clear neutron flux, but are very consistent with the calibration data used.

Sample	Live (s)	Mass (g)	Cd114	Bq/kg	Mn56	Bq/kg	Zn63	Bq/kg	Na24	Bq/kg	K42	Bq/kg	Al28	Bq/kg	Mn56	Bq/kg	Na24	Bq/kg
peak [keV]			558.5		846.76		962.07		1368.5		1524.8		1779		1810.3		2754
Yield (%)			100		98.85		6.53		99.99		18.08		100		26.9		99.87
efficiency (%)			3.3547		2.5381		2.2702		1.7219		1.5122		1.3516		1.3444		0.9783
MN 270 01 N	37,300	77.88	0		27,538	377.375	397	92.1982	54713	1093.78	1040	130.993	Ld	Ld	4109	391.251	31,268	1101.71
MN 270 06 N	45,705	78.37	0		28,882	320.982	257	48.4035	56787	920.659	1002	102.352	Ld	Ld	4281	330.58	32,321	923.556
MN 270 07 N	33,024	112.3	178	1.43	41,891	449.663	501	91.1371	82742	1295.65	1501	148.089	147	2.93616	6084	453.788	46,810	1291.91
MN 270 20 N	52,273	74.18	0		38,051	390.638	266	46.2787	67785	1015.17	1688	159.278	412	7.87049	5516	393.47	38,960	1028.38

) did not evidence many of the expected energies, with exceptions being made for ¹¹⁴Cd, ⁶³Zn, ⁵⁶Mn, ⁵⁶Fe, and ²⁸Al. However, some other gamma lines were identified for ⁴²K and ²⁴Na, for example, which is explicable. There is a remarkable equivalence of the activities found for the different energies of ²⁴Na and ⁵⁶Mn, showing that the calibration and simulation processes are valid and precise. The isotopic activities are compatible with the elemental analysis performed by XRF, but given the lack of information on the precise flux and energy of the neutrons, it is hazardous to define a specific activity with associated uncertainty for the activation process.

3.3. Radiometric Results in the Multi-Proxy Context

Natural-series radionuclides measured by HPGe fall within background ranges for surface shelf sediments. ²²⁸Ac activities systematically exceed ²⁰⁸Tl, consistent with ²²⁰Rn loss in near-surface matrices of the ²³²Th chain. ⁴⁰K is ubiquitous and reflects the abundance of K-bearing aluminosilicates. Spatially, ⁴⁰K co-varies with the terrigenous carriers indicated by Fe-Ti-V (EDXRF). Anthropogenic ¹³⁷Cs is still detectable at several stations, decades after Chernobyl, whereas ²⁴¹Am and ²³⁹Pu were below the detection limits. Because the 186 keV line conflates ²³⁵U and ²²⁶Ra, and the 144 keV assumption of secular equilibrium in the ²³⁵U series was not met for all samples, U-series activities are reported conservatively as screening-level values.

Short-irradiation neutron-activation checks detected expected short-lived products (⁵⁶Mn, ⁵⁶Fe, and ²⁸Al) and matrix lines (²⁴Na, ⁴²K); the close agreement between multiple γ lines of the same isotope (e.g., ²⁴Na, ⁵⁶Mn) supports calibration fidelity. The activation pattern is consistent with bulk composition from EDXRF. In the absence of a quantified neutron flux/energy spectrum, values in Table 14 are provided as qualitative corroboration rather than absolute activities.

γ-proxies align with PCA/FA structure: ^{^40}K co-varies with Fe-Ti-V (terrigenous), whereas ^{^137}Cs co-occurs at stations with elevated Pb/Hg and HMW-PAHs (anthropogenic). These radiometric markers independently corroborate the three-axis solution.

4. Discussion

Coastal sedimentary environments along the Romanian shelf are shaped by well-known drivers—riverine loading, nearshore anthropogenic pressures, hydrodynamic sorting, seasonal stratification, and early diagenesis—but the value of the present dataset is that it resolves how these controls co-vary across water-sediment compartments and alongshore–cross-shore gradients at the same time. Two “new” outcomes emerge from this integrated approach: a coherent mechanistic separation between basin-scale Danube modulation and local coastal hotspots, visible simultaneously in nutrients, organic matter, dissolved metals, sediment geochemistry, and multivariate structure; and a consistent fingerprint of matrix control (Fe-oxide/fine terrigenous carriers and carbonate dilution) that explains why some apparent “contaminant signals” (e.g., Ni, Cr) can be largely lithogenic, whereas others (e.g., Pb, Hg, and selected PAHs) present localized nearshore enrichment. Surface nutrients remain spatially heterogeneous, with the highest values in the Danube Delta influence area and secondary coastal hotspots in the south. In the Romanian shelf waters, mean surface concentrations were 0.15 ± 0.09 µM for phosphate, 1.33 ± 7.74 µM for silicate, 0.37 ± 0.22 µM for nitrate, 0.03 ± 0.05 µM for nitrite, and 1.20 ± 1.16 µM for ammonium. The unusually large variability in silicate (±7.74 µM) reflects the strong contrast between the Danube plume and rapidly diluted, biologically depleted shelf waters [73,74,75]. This dichotomy is essential for interpreting nutrient status: high northern concentrations (particularly SiO₄⁴⁻, NO₃⁻, and PO₄³⁻) are primarily controlled by diffuse Danube inputs [74,75], making the delta-influenced zone the main driver of nutrient fields on the northern shelf [76]. In contrast, southern nearshore enrichments are better explained by coastal point sources [76,77], producing nutrient-specific behavior. Silicate tracks terrestrial weathering and is therefore tightly coupled to river dominance [78,79]; the terrigenous sediment supply typical of deltaic settings [78] provides a consistent explanation for elevated dissolved Si in the north. Against this background, the most diagnostically “new” aspect of the present nutrient dataset is the juxtaposition of reduced oxidized nutrients (NO₃⁻, PO₄³⁻) with persistently elevated NH₄⁺, which indicates that diffuse-source mitigation can coexist with strong localized N recycling and organic matter turnover. Phosphate in the present study is slightly higher than the pre-eutrophication baseline (0.11 ± 0.04 µM before the 1970s [80]) but far below the mid-1970s peak (1.34 ± 0.33 µM) and the 2006–2011 period (0.31 ± 0.96 µM) [81]. Nitrate shows an even clearer multi-decadal decline, from 15.8 ± 9.4 µM in the mid-1970s and 5.18 ± 6.73 µM in 2006–2011 to 0.39 ± 0.23 µM here. Nitrite and ammonium are also lower than in 2006–2011 (0.89 ± 3.11 µM and 4.15 ± 9.42 µM, respectively [81]), consistent with the transition toward reduced eutrophication pressure documented for the mid-2000s and onward [79], and with regional assessments using nutrient dynamics and eutrophication indices [82,83]. Silicate in the present dataset is far below mid-1960s values (41.3 ± 7 µM) and below those reported for 1985–1993 (6.0 ± 2.0 µM), indicating a major long-term shift in either Danube delivery, dilution/uptake, or spatial coverage of sampling. Importantly, the higher nutrient concentrations reported in shallow Constanta waters during 2020–2023 [84] likely reflect differences in spatial and temporal representativeness (single-station, daily observations) rather than a direct contradiction—highlighting why multi-station surveys and fixed-point monitoring can yield different “typical” values, even within the same coastal sector. The basin context further clarifies what is distinctive. Surface phosphate values here are similar to those reported for the North and Northeastern Black Sea (0.05–0.3 µM and 0.1–0.2 µM [85,86]) and lower than values reported for the Southeastern Black Sea (0.34 ± 2.23 µM [87]). Surface silicate is comparable to the Northern Black Sea (0.5–2.7 µM [85]) and slightly lower than northeastern/southeastern coastal waters (approx. 2–3 µM [86,87]). Nitrate resembles reported coastal values in the northeastern and southeastern basin (0.27–0.70 µM and 0.27 ± 0.83 µM [87,88]). The major contrast is ammonium: surface NH₄⁺ here is higher than the Northern Black Sea average (0.19–0.22 µM), reinforcing the conclusion that local recycling and/or nearshore sources remain important, even as oxidized nutrients decrease. Elevated NH₄⁺ is consistent with rapid regeneration during organic matter remineralization following intense productivity in nutrient-rich embayments (e.g., Portita Bay) [75,89]. While earlier data suggested continuous anthropogenic point-source inputs (e.g., wastewater, urban runoff) [77], the present observations do not consistently support that mechanism, implying improved wastewater treatment effectiveness in key hotspots and shifting the dominant control toward internal cycling and localized organic loading rather than direct NH₄⁺ discharge [90].

4.1. Hydrography, Nutrients, and Organic Matter: Danube Forcing, Coastal Regeneration, and Interannual Context

The observed vertical structure—bottom maxima of PO₄³⁻, SiO₄⁴⁻, and NH₄⁺, a subsurface chlorophyll-a maximum at 12–16 m, and slightly higher TN at depth—is characteristic of seasonally stratified inner-shelf waters in the Northwestern Black Sea. It reflects remineralization and benthic release beneath the euphotic layer, coupled with an efficient surface uptake of oxidized nitrogen species (NO₃⁻/NO₂⁻). In our profiles, NH₄⁺ dominated inorganic N at most stations (up to 19.33 µM at S23/16 m), whereas NO₃⁻ and NO₂⁻ peaked below the chlorophyll maximum at S13 (15–25 m). This partitioning is consistent with summertime nutrient drawdown in surface waters and enhanced regeneration at depth, and it aligns with regional syntheses describing summer stratification, near-bottom nutrient accumulation, and river-modulated surface properties on the Romanian shelf [91,92]. The coupling of organic matter indicators with nutrient structure provides a more process-based interpretation than nutrients alone. Surface TOC enrichments in the northern sector (e.g., 4.85 mg L⁻¹ at S01 and 4.09 mg L⁻¹ at S23) support the presence of allochthonous organic inputs associated with Danube-influenced waters, superimposed on in situ production. Meanwhile, the persistence of high NH₄⁺—including the extreme value at S23—indicates that regeneration can dominate the inorganic N pool locally, even when NO₃⁻ remains low. Relative to earlier periods characterized by elevated oxidized nutrients across large shelf areas [79,81], the present vertical patterns suggest a system increasingly governed by reduced diffuse oxidized-N and phosphate loading, and spatially patchy internal recycling, particularly in nearshore retention zones.

4.2. Dissolved Metals in Surface Waters: Nearshore Enrichment on a Low-Hg Background

Dissolved trace metals showed gradients consistent with nearshore inputs and/or resuspension: Fe = 1.12–5.91 ppb (maximum at S06), Zn = 4.25–10.77 ppb (maximum at S01), and Cu < 0.45–38.37 ppb (maximum at S06). Potentially toxic elements were elevated mainly at shallow stations (Pb up to 27.4 ppb at S01; Cr up to 4.04 ppb at S20; As up to 3.85 ppb at S01; Cd up to 1.32 ppb at S25; and Co up to 1.63 ppb at S06), consistent with harbor/urban influences and wave- or traffic-induced resuspension. In contrast, dissolved Hg remained uniformly low (0.03–0.05 ppb), suggesting limited recent inputs and/or efficient scavenging under prevailing particle/redox conditions [93]. The spatial organization and magnitude of these concentrations are consistent with recent Romanian-sector assessments used for benchmarking hazardous substances against national quality frameworks [16,94]. A key inter-period comparison is that our dissolved concentrations largely fall within the variability reported for 2006–2011 marine waters (e.g., Cu 10.02 ± 13.77 µg/L; Cd 0.99 ± 1.26 µg/L; Pb 3.78 ± 6.03 µg/L; Ni 3.65 ± 4.40 µg/L; and Cr 3.84 ± 6.26 µg/L; with broad ranges) [95]. Two differences are nonetheless notable. First, our dataset shows localized Cu and Pb highs (e.g., Cu at S06, Pb at S01), consistent with point-influenced nearshore settings rather than offshore gradients. Second, our reported Ni is frequently below the stated detection threshold (<0.85 ppb) at several stations, which contrasts with the multi-year means from earlier monitoring. This discrepancy is plausibly explained by a combination of true spatial heterogeneity (few stations, nearshore bias) and analytical/detection-limit constraints relative to long-term monitoring programs. Overall, the dissolved-metal results support the interpretation that present-day pressures are dominated by nearshore episodic/point influences superimposed on a background where Hg inputs are currently limited.

4.3. Surface-Sediment Geochemistry: Terrigenous Control, Carbonate Dilution, and Localized Hotspots

Contrary to the canonical offshore CaCO₃ increase sometimes reported for inner shelves, our dataset shows higher CaCO₃ at the shoreface (e.g., S01 = 28.74%, S05 = 28.04%, and S06 = 29.87%) and lower CaCO₃ offshore (e.g., S18 = 10.09%, S19 = 7.60%, and S20 = 7.96%). This cross-shore pattern is consistent with bioclastic enrichment in the surf/shoreface (shell debris accumulation) and progressive dilution by siliciclastic material seaward. Sedimentary TOC is low overall (0.11–0.96%; mean 0.45%) and remains ≤1% at all stations, with patchy enrichments both nearshore (S10 = 0.779%, S12 = 0.785%) and offshore (S24 = 0.964%). Such patchiness is typical of environments where grain size exerts strong control (fines fractions OM preservation) and where episodic resuspension/settling and local hydrodynamics redistribute organic matter [7]. These facies constraints are important when comparing metal loads across studies because Romanian monitoring programs repeatedly show that finer, TOC-richer sediments retain higher metal inventories than coarse nearshore sands [95,96,97].

4.3.1. Metal Carriers and Source Context

The coastal zone is heavily anthropized and exposed to multiple metal sources, including Danube-borne domestic/municipal/agricultural/industrial effluents, and local inputs linked to ports, shipyards, coastal tourism, construction, petrochemical activity, wastewater discharges, and offshore oil and gas operations. Agricultural runoff (fertilizers/pesticides) can elevate As, Cu, Ni, Pb, and Zn [95,96,97,98,99,100,101,102,103]. Untreated or poorly treated wastewater is an established contributor of As, Cr, Cu, Hg, Pb, and Zn [98,101,103,104,105,106], while waterborne transport pathways are often invoked for Ni, Pb, and Zn [98,103,105,106]. Anti-fouling coatings are a recognized anthropogenic source of Cu and Zn [98,102], and offshore platforms/refineries/construction can supply As, Cr, Cu, Hg, Ni, Pb, and Zn via wastewater discharge and atmospheric deposition [101,102,107]. In parallel, natural weathering/erosion contributes terrigenous trace-element (TE) fluxes, and many studies emphasize that clay, silt, and organic particles are key TE carriers [8,105,106,108,109,110,111,112,113,114,115,116,117]. Several papers also report natural components for Cr, Hg, Pb, and Zn in sediments [98,103,106,110,111,112,117,118,119,120,121,122,123]. Our data directly support a matrix-dominated (terrigenous) host phase for most trace metals. The correlation matrix shows strong Fe covariation with V (r = 0.927), Ni (0.945), Cu (0.819), Zn (0.921), and Pb (0.872), plus a moderate Fe-Cr link (0.598), indicating a dominant role of fine aluminosilicates and Fe-oxide/hydroxide coatings as carriers. Chromium maxima cluster at shallow Eforie stations (S08 = 121 ppm; S10 = 101 ppm; S15 = 102 ppm; S16 = 116 ppm; S25 = 120 ppm) and correlate with Ti (Ti–Cr r = 0.651), consistent with a lithogenic contribution enriched in heavy-mineral/oxide phases.

4.3.2. Concentrations in This Study and Comparison with Romanian Monitoring over Time

It should be noted that, in dry sediments, 1 µg g⁻¹ is approximately equivalent to 1 ppm. Therefore, the concentrations reported here in ppm are directly comparable with the µg g⁻¹ values commonly used in Romanian monitoring reports. In the MN 270 surface-sediment dataset, chromium spans 63.3–121 ppm (mean 94.08 ppm), nickel 26.65–52.10 ppm (mean 38.88 ppm), copper 7.35–38.53 ppm (mean 20.10 ppm), lead 13.09–25.66 ppm (mean 19.63 ppm), zinc 41.77–89.00 ppm (mean 64.18 ppm), arsenic 4.88–10.90 ppm (mean 7.33 ppm), and mercury 0.016–0.090 ppm (mean 0.06 ppm). With the exception of nickel, the arithmetic means for As, Cr, Cu, Pb, Hg, and Zn remain below Romanian sediment-quality criteria [20]. Nickel, however, averages 38.88 ppm, i.e., slightly above the 35 ppm threshold. Considering the strong covariation of Ni with Fe (r = 0.945) and V (r = 0.982), together with the regional lithological context, this minor exceedance is more plausibly attributed to natural enrichment and adsorption/coprecipitation with Mn/Fe oxides, rather than a discrete point-source input [74,124,125].

Long-term national monitoring provides a useful benchmark for temporal context. For the 2006–2011 period, Oros (2019) [95] reports mean ± SD values (and ranges) in superficial marine sediments of Cu = 33.05 ± 27.54 µg g⁻¹ (0.53–147.84 µg g⁻¹), Cd = 1.03 ± 1.52 µg g⁻¹ (0.01–9.63 µg g⁻¹), Pb = 26.71 ± 26.16 µg g⁻¹ (0.10–300.78 µg g⁻¹), Ni = 36.54 ± 25.93 µg g⁻¹ (0.40–211.73 µg g⁻¹), and Cr = 44.58 ± 31.08 µg g⁻¹ (1.34–231 µg g⁻¹). Relative to these 2006–2011 means, our Cu (mean 20.10 ppm) and Pb (mean 19.63 ppm) are lower, while Ni (mean 38.88 ppm) is very close to the 2006–2011 mean (36.54 µg g⁻¹). The most pronounced difference concerns Cr, for which our mean (94.08 ppm) is substantially higher than the 2006–2011 mean (44.58 µg g⁻¹). This discrepancy is consistent with the strong lithogenic control observed in our dataset (including Ti-Cr coupling) and indicates that local heavy-mineral/oxide enrichment can dominate Cr distributions, even when other metals remain moderate. For the 2012–2017 assessment, Oros (2019) [95] summarizes MSFD-style evaluations using percentile statistics, showing that Ni frequently failed GES across reporting units, with the 75th percentile values markedly above the 35 µg g⁻¹ threshold (e.g., marine waters sediments: Ni P75 = 78.09; coastal waters: 44.15; variable-salinity waters: 87.47, all versus threshold 35). In the same framework, Cu showed poor status in some units (marine waters sediments: Cu P75 = 51.08 vs. threshold 40; variable-salinity waters: 44.25 vs. 40), whereas Cr and Pb generally remained below ERL-style thresholds (e.g., Cr threshold 81; Pb threshold 47). Against this background, our mean Ni = 38.88 ppm is only slightly above 35 and is therefore far below the Ni enrichment captured by the 2012–2017 75th percentile values, which is consistent with the expectation that the strongest Ni signals are often expressed in fine-grained depositional sinks rather than in coarser nearshore substrates. For 2018, Oros (2019) [95] reports average concentrations in superficial marine sediments of Cu = 24.57 ± 12.98 µg g⁻¹, Cd = 0.25 ± 0.29 µg g⁻¹, Pb = 10.43 ± 6.09 µg g⁻¹, Ni = 72.10 ± 27.59 µg g⁻¹, and Cr = 23.92 ± 10.48 µg g⁻¹. Compared with these 2018 averages, our dataset shows substantially lower Ni (38.88 vs. 72.10 µg g⁻¹) but higher Pb (19.63 vs. 10.43 µg g⁻¹) and much higher Cr (94.08 vs. 23.92 µg g⁻¹). This divergence aligns with Romanian monitoring interpretations that deeper bathymetric zones (30–100 m) and fine sediments tend to concentrate Ni, whereas sector-specific nearshore facies and heavy-mineral inputs can elevate Cr [95,126]. More recently, the 2023 Romanian shelf investigation reported by Oros et al. (2025) [94] provides averages and ranges of Cu = 11.263 ± 10.882 µg g⁻¹ (1.622–61.140), Cd = 0.414 ± 0.595 µg g⁻¹ (0.053–2.837), Pb = 14.364 ± 8.930 µg g⁻¹ (3.095–43.160), Ni = 14.765 ± 6.760 µg g⁻¹ (5.431–28.730), Cr = 11.328 ± 4.691 µg g⁻¹ (4.962–24.260), and Co = 2.896 ± 0.838 µg g⁻¹ (1.403–4.815). Our mean values for Cu (20.10 ppm), Pb (19.63 ppm), Ni (38.88 ppm), and especially Cr (94.08 ppm) are higher than these 2023 averages, and our Ni and Cr ranges exceed the 2023 maxima reported for those elements. Importantly, this contrast should not be interpreted automatically as a straightforward deterioration of sediment quality. A more defensible explanation is that methodological and sampling/facies differences—including station distribution (nearshore heavy-mineral/oxide-rich sediments versus finer offshore muds), depth-window selection, and potential differences in digestion/extraction protocols or targeted grain-size fractions—can materially shift the “total” concentrations reported. This interpretation is consistent with the long-standing Romanian monitoring conclusion that variability in Cr, Cu, Ni, and Pb depends strongly on element behavior, sediment texture, TOC, water depth, and mixed anthropogenic and natural controls [95,96,97].

4.3.3. Black Sea Regional Context and the Ni “Background” Issue

Regional studies support the view that Romanian shelf sediments frequently show comparatively high Ni relative to Bulgarian and Turkish areas. Simeonov et al. (2000) [127] reported higher mean As and Cu than in our study but lower Ni, and interpreted Cu as mainly natural, while As, Cr, Ni, Pb, and Zn had stronger anthropogenic contributions near industrialized Bulgarian sectors. Turkish-coast studies likewise often report Cu, Pb, and Zn higher than in our dataset, but Ni lower [107,128,129,130]. This basin-scale contrast is consistent with the observation that the natural geochemical value for Ni in NW Black Sea sediments was approx. 5.9 mg/kg in the 1980s but is currently >40 mg/kg [131]. Therefore, the slight exceedance of the 35 mg/kg guideline in our dataset (mean Ni = 38.88) should be interpreted cautiously, in line with the repeated recommendation that Ni thresholds may be below the local background in parts of the Black Sea [95,126].

4.4. PAHs in Water and Sediments: Petrogenic Dominance with Local Pyrogenic Pulses

Σ₁₆PAH in surface waters (134–347 ng L⁻¹; highest at S01/S06/S07) and in surface sediments (42–88 µg kg⁻¹; maximum at S06) place the investigated sites in a low-to-moderate regional band. In both compartments, mixtures are skewed toward low-molecular-weight (LMW, 2–3 ring) compounds (notably naphthalene, acenaphthene, and fluorene), supporting a predominantly petrogenic signature related to fuel/oil handling and maritime traffic. At the same time, localized enrichments of higher-molecular-weight (HMW) markers—benzo[g,h,i]perylene in water at S01, and indeno(1,2,3-cd)pyrene and dibenzo[a,h]anthracene in sediments at S01 and S13—indicate episodic pyrogenic contributions (combustion aerosols, shipping exhaust) superimposed on the petrogenic background. Although diagnostic ratios (Ant/(Ant + Phe), Fl/(Fl + Py), BaA/(BaA + Chry)) are widely used to apportion sources, they can be altered by weathering/aging. Within that constraint, our patterns support mixed inputs dominated by petrogenic PAHs with localized pyrogenic pulses [91]. A temporal comparison suggests that the present PAH burden is generally lower than recent Romanian monitoring maxima. In 2023, PAHs in Romanian shelf sediments ranged up to 0.428 µg/g dry weight (i.e., 428 µg/kg), with most individual compounds near detection limits, except fluorene, anthracene, and indeno(1,2,3-c,d)pyrene, which dominated particularly in the Danube-influenced area [94]. Our Σ₁₆PAH maxima (≤88 µg/kg) are well below that reported maximum, indicating either lower local accumulation in the sampled sectors and/or differences in the spatial targeting of high-risk depositional environments (fine-grained, organic-rich sinks) emphasized in broader monitoring. In parallel, the multivariate structure adds a mechanistic element not available from concentration maps alone: in the PAH-augmented PCA (six stations), Σ₁₆PAH co-varies with Fe-oxide-associated metals (e.g., Ni, Zn, Pb, and Cu), consistent with transport and retention on fine terrigenous carriers and supporting the interpretation of matrix-controlled accumulation with localized source pulses. Finally, the wider literature confirms that the Romanian sector is predisposed to mixed PAH inputs. PAHs are predominantly produced by the combustion of carbonaceous materials [132], but can also derive from petrogenic releases and natural processes [133]. In the Black Sea, both pyrolytic and petrogenic patterns have been reported, varying by coastline sector [10]. Romanian monitoring has repeatedly identified PAH compounds of concern (e.g., fluorene, phenanthrene, anthracene, benzo(a)pyrene, benzo(g,h,i)perylene, and dibenzo(a,h)anthracene) in relation to maritime traffic and offshore operations [94,134,135], consistent with the localized HMW signals observed here. The present dataset, therefore, refines basin-wide understanding by showing that, even under generally low-to-moderate totals, pyrogenic pulses can be detected at specific nearshore stations while the regional background remains petrogenic—an interpretation strengthened by the coupled water–sediment measurements and the PCA-based carrier structure.

4.5. Radiometric Constraints: Role and Limits in Coastal Matrices

HPGe γ-spectrometry of surface sediments adds an independent constraint to the multi-proxy interpretation by quantifying natural-series radionuclides that principally track detrital mineral phases and by detecting particle-reactive anthropogenic tracers, most notably ^{^137}Cs. In our MN 270 dataset, the γ-spectra indicate background-level activities for the natural radionuclides, consistent with a dominantly lithogenic matrix. Within the ^{^232}Th decay chain, ^{^228}Ac activities systematically exceed ^{^208}Tl, a pattern that is typical for near-surface sediments where the gaseous loss of ^{^220}Rn can partially decouple the lower part of the chain. In parallel, ^{^40}K is ubiquitous and behaves as a first-order proxy for K-bearing aluminosilicates (e.g., illite and feldspars), providing a radiometric expression of detrital supply that is coherent with the terrigenous control inferred from major and trace elements.

This interpretation is strengthened by the close agreement between radiometric and geochemical structure in our dataset. The spatial behavior of ^{^40}K mirrors the variability of Fe-Ti-V derived from EDXRF and supports the same “Terrigenous” axis identified by PCA/FA, in which trace metals covary with aluminosilicates and Fe-oxide carrier phases. Consistently, the ^{^232}Th-series signature (including the characteristic ^{^228}Ac > ^{^208}Tl relationship) is compatible with fine detrital phases acting as the dominant host for trace metals, reinforcing the conclusion that the metal distribution is largely governed by sediment mineralogy and texture rather than by uniformly elevated contaminant inputs. These observations align with widely applied international practices for using natural-series nuclides and ^{^137}Cs as complementary indicators in coastal and shelf monitoring frameworks [70,71,72,73]. Against this broader regional backdrop, our results are also consistent with Romanian and Black Sea sediment inventories reported in the literature, which generally show a wide but “background-dominated” envelope for natural radionuclides and a spatially heterogeneous, yet persistent, ^{^137}Cs signal reflecting post-Chernobyl fallout retention in sediments. For example, Chiroșca et al. 2018 [136] documented broad activity ranges in Romanian shelf surface sediments, ^{^40}K = 77–267 Bq kg⁻¹, ^{^232}Th = 3.51–88.10 Bq kg⁻¹, ^{^226}Ra = 3.56–66.40 Bq kg⁻¹, and ^{^137}Cs = 0.94–26.4 Bq kg⁻¹, with higher values commonly associated with a stronger terrigenous influence and/or fine-grained depositional settings that favor particle retention and inventory focusing. Within this Romanian shelf context, the MN 270 radiometric imprint conforms to the same “natural background plus anthropogenic marker” pattern: natural-series activities reflect lithogenic control, while ^{^137}Cs, where present, indicates an externally supplied anthropogenic component superimposed on the detrital matrix. The distribution of ^{^137}Cs in our dataset also provides a useful interpretive complement to the chemical indicators of local pressure. Because ^{^137}Cs is strongly particle-reactive, it is preferentially retained where fine sediments and organic matter promote sorption and accumulation, whereas mobile, sand-dominated substrates tend to dilute and redistribute such tracers. This sedimentological control helps explain why cross-study differences in measured activities (and, by extension, in apparent “hotspots”) should not be interpreted as straightforward temporal deterioration or recovery unless sampling design and sedimentary facies are directly comparable. In our study area, characterized by pronounced facies heterogeneity and nearshore reworking, the radiometric signal therefore primarily constrains provenance and depositional focusing, while still offering an independent line of evidence for spatially variable anthropogenic influence through the persistence of ^{^137}Cs. Finally, the absence of a detectable transuranic imprint in our samples (i.e., ^{^241}Am and ^{^239}Pu below detection limits) is compatible with the general Romanian shelf perspective that these nuclides are not consistently resolved in surface sediments at the screening level and are strongly modulated by sediment texture, mixing, and depositional focusing. Overall, the combined radiometric signature, ^{^40}K and Th/Ra-series nuclides tracing the detrital background, and ^{^137}Cs marking anthropogenic influence mirrors the terrigenous/anthropogenic separation found in PCA/FA and strengthens the causal interpretation of our multi-proxy results beyond simple co-occurrence [137,138,139,140].

4.6. Synthesis: A Three-Axis Environmental Signature Supported by Correlations

The multivariate structure of the sediment dataset resolves three persistent and interpretable dimensions of variability, which are mutually consistent across geochemical, organic, and radiometric indicators. The first dimension is a terrigenous (detrital–mineral) axis, defined by strong and coherent covariation among matrix elements and associated trace metals (Fe-Al-Ti-V with Ni and, more moderately, Cr). This association, together with the close correspondence between these elements and the radiometric behavior of ^{^40}K and the Th-series nuclides, indicates that trace metals are predominantly hosted by the fine detrital fraction and Fe-oxide/hydroxide carrier phases rather than occurring as independent contaminant enrichments. In this framework, ^{^40}K provides an independent mineralogical proxy for K-bearing aluminosilicates, reinforcing the interpretation that the dominant control on metal inventories is lithogenic and texture-dependent. The second dimension is a carbonate–facies axis, in which CaCO₃ (and Sr, where available) captures contrasts between bioclastic and siliciclastic sedimentary domains. This facies control is expressed by elevated CaCO₃ values in the shoreface, where shell debris and biogenic carbonate are more abundant, and lower CaCO₃ offshore, where carbonate is diluted by siliciclastic inputs. Importantly, this axis is largely decoupled from the terrigenous carrier axis, indicating that carbonate dilution and bioclastic enrichment act as an independent control on bulk sediment composition and on the apparent concentrations of matrix-bound constituents. The third dimension is an anthropogenic (nearshore pressure) component, expressed most clearly by the co-occurrence of elevated Pb and Hg in sediments with localized increases in high-molecular-weight PAHs, and, where present, the particle-reactive fallout tracer ^{^137}Cs. This component explains small-scale departures from the dominant lithogenic and facies gradients and is spatially coherent with the nearshore zone, where urban/port influences and coastal reworking are most pronounced. In the PCA configuration, nearshore stations tend to project toward the vectors associated with this anthropogenic signal, whereas stations dominated by fine detrital matrices align along the terrigenous direction. The stability of these three axes across chemical and radiometric proxies supports a robust separation between detrital/mineral control, carbonate facies effects, and localized anthropogenic imprints in the Romanian sector of the Black Sea.

4.7. Limitations and Outlook

Three considerations constrain the strength of source apportionment and the extent to which the present results can be generalized beyond this survey. First, the temporal scope is limited to a single early-summer campaign; seasonal variability, particularly the effects of autumn/winter destratification and mixing, changes in Danube influence, and episodic storm-driven resuspension, was not resolved. Second, the spatial resolution and analytical coverage are uneven across proxies, with PAHs and several contaminant-focused indicators quantified at a limited subset of stations; consequently, narrow coastal gradients and small-scale port-related hotspots may be under-represented relative to basin-scale facies controls. Third, normalization and covariate control were not applied systematically: metals were not normalized to grain size or black carbon, and carbonate dilution was not explicitly corrected, which can reduce cross-facies comparability and complicate quantitative separation between matrix control and superimposed anthropogenic inputs. Addressing unresolved seasonality requires sampling that captures the dominant hydrographic regimes that structure coastal biogeochemistry in the Northwestern Black Sea. A practical next step is a minimum two-season design targeting late autumn or winter conditions that are representative of vertical mixing and enhanced sediment remobilization, and spring/early summer conditions representative of strong riverine forcing and stratification onset, complemented, where feasible, by event-based sampling after major storms or high-discharge periods. This design would allow direct testing of whether the terrigenous, carbonate, and anthropogenic axes identified here are stable across seasons or whether the relative contributions of benthic release, resuspension, and nearshore point sources shift under mixed-water-column conditions.

The limitation of spatial resolution can be evaluated more directly through structured comparison with Romanian monitoring programs and sectoral studies that employ denser station networks along the coast. These datasets consistently show that contaminant variability is strongly modulated by sediment texture, TOC, and depth, while localized maxima occur near major harbors and urbanized stretches. Therefore, differences between studies often reflect station distribution and sampled facies rather than true temporal change. In this context, the MN 270 grid captures the dominant cross-shelf facies gradient but may smooth sharp nearshore discontinuities. A refined design for future work would densify station spacing in high-pressure areas (e.g., harbor approaches and wastewater-affected sectors) and apply stratified sampling by sediment class (sand vs. muddy sand vs. mud) to ensure that hotspot detection is not confounded by facies-driven dilution.

Future work should therefore prioritize expanded PAH characterization (including alkylated homologues and complementary diagnostic approaches used alongside LMW/HMW and co-tracers), grain-size and black-carbon normalization to better separate texture-driven enrichment from true source effects, and the incorporation of independent sediment-dynamic constraints (e.g., excess ²¹⁰Pb) to quantify sediment focusing and provide sedimentation-rate context. In the interim, and for management benchmarking, the sediment concentrations measured here remain generally low relative to Romanian regulatory criteria [20], while the slight Ni elevation is consistent with the well-documented role of regional lithology and mineral–oxide associations in shaping “background” Ni levels.

4.8. Mechanistic Interpretation of the Multivariate Structure

The integrated proxy evidence links the multivariate structure to plausible sedimentological and biogeochemical mechanisms. The dominant component reflects mineral–matrix control on trace metals, with Fe oxyhydroxides and aluminosilicate-rich detrital fractions acting as the primary host phases. This carrier control explains the coherent covariation of Fe–Ti–V with associated trace metals, particularly Ni (and, more moderately, Cr), consistent with adsorption and co-precipitation onto fine terrigenous particles and oxide coatings. A second, largely independent pattern is expressed by CaCO₃ (and Sr where available) and captures facies-scale contrasts between carbonate-enriched shoreface bioclastic settings and offshore deposits increasingly diluted by siliciclastics. The weak coupling between this carbonate signal and the detrital metal–carrier axis indicates that facies-driven dilution/enrichment modulates bulk composition without necessarily controlling the trace-metal inventory. Superimposed on these natural gradients, a distinct nearshore anthropogenic imprint is indicated by the spatial increase of Pb, Hg, and high-molecular-weight PAHs toward coastal sectors influenced by urban and port activities; where detectable, this imprint is further supported by the co-occurrence of ^137Cs. Importantly, these contaminant-associated variables remain separable from mineral–matrix proxies (e.g., Fe) in multivariate space, indicating that localized inputs and nearshore depositional/reworking processes generate spatially restricted enrichments even when the background is strongly governed by lithogenic carriers and facies variability. Overall, the results show that most variability in the geochemical dataset is controlled by terrigenous mineral matrices and carbonate–siliciclastic facies contrasts, while anthropogenic signals are detectable but spatially confined and secondary to the dominant natural controls in this dynamically reworked inner-shelf environment.

5. Conclusions

The MN 270 survey documents an inner-shelf system in which summer hydrography, Danube-modulated material supply, and strong lithofacies contrast jointly in structure with the distribution of nutrients, trace metals, organic contaminants, and radiometric markers.

The water-column hydro-biogeochemical regime was characteristic of the warm season, with stratification expressed by the bottom-layer maxima of PO₄³⁻, SiO₄⁴⁻, and NH₄⁺ and a pronounced deep chlorophyll-a maximum at 12–16 m. Patterns in TOC and TN, including higher surface TOC in the northern sector, indicate predominantly in situ production superimposed on a measurable allochthonous contribution consistent with Danube influence.

In surface waters, essential dissolved metals (Fe, Zn, and Cu) displayed localized nearshore enrichments, while potentially toxic metals (Pb, Cr, As, Cd, and Co) tended to peak at shallow stations, consistent with harbor/urban pressures and resuspension-driven mobilization. Dissolved Hg remained uniformly low (0.03–0.05 ppb), suggesting limited recent inputs and/or effective removal from the dissolved phase under prevailing coastal conditions.

The surface-sediment record confirms a shoreface-to-offshore facies gradient. CaCO₃ is highest in the shoreface and decreases offshore (7.6–29.9%, mean 13.2%), reflecting bioclastic enrichment nearshore and progressive siliciclastic dilution seaward. Sedimentary TOC is uniformly low (0.11–0.96%, mean 0.45%), consistent with well-aerated, sand-dominated substrates and generally limited organic matter preservation, with patchy enrichments where fine fractions accumulate.

Trace-metal distributions are primarily controlled by mineral carriers and sediment texture. Strong Fe-centered associations with V, Ni, Cu, Zn, and Pb (and a moderate association with Cr) indicate that metals are largely hosted by aluminosilicate-rich detritus and Fe-oxide/hydroxide phases, supporting a predominantly lithogenic control. Elevated Cr at shallow Eforie stations is consistent with terrigenous inputs enriched in oxide/heavy-mineral fractions and illustrates how local mineralogical variability can dominate Cr patterns, even where other elements remain moderate.

In the regulatory context, arithmetic means for As, Cr, Cu, Pb, Hg, and Zn are below Romanian sediment-quality criteria (Order 161/2006). Nickel averages 38.88 ppm, slightly exceeding the 35 ppm threshold. However, its strong covariance with Fe and V and the known regional geological setting support interpretation as natural enrichment linked to lithogenic sources and Fe/Mn-oxide association rather than widespread point-source contamination.

Although natural controls dominate, localized anthropogenic imprints are evident. The nearshore maxima of Pb and Hg (e.g., at S23) and their co-enrichment with Fe and Zn indicate coastal inputs related to harbor/urban activity superimposed on natural carrier phases.

PAH concentrations in both waters and sediments fall within a low-to-moderate range and are dominated by low-molecular-weight compounds, consistent with a petrogenic background linked to maritime activity. Station-specific high-molecular-weight spikes (e.g., benzo[g,h,i]perylene in water; indeno(1,2,3-cd)pyrene and dibenzo[a,h]anthracene in sediments) indicate episodic pyrogenic contributions attributable to combustion-related aerosols and exhaust.

The radiometric inventory derived from HPGe γ-spectrometry provides a complementary baseline in sediments that is fully coherent with the elemental and organic proxy interpretation and does not indicate any radiological anomaly relative to the expected shelf background.

Finally, the multivariate synthesis integrates these observations into three near-orthogonal environmental controls: a terrigenous axis (Fe-Al-Ti-V with Ni and Cr), a carbonate axis (CaCO₃, with Sr where available), and an anthropogenic factor (Pb, Hg, and HMW-PAHs). Lithofacies contrasts establish the dominant background variability, onto which localized contaminant signals are superimposed.

Overall, the results depict a Danube-influenced, shoreface-dominated inner shelf where lithogenic and facies gradients primarily regulate trace-metal distributions, while discrete nearshore hotspots capture anthropogenic inputs. This evidence base supports targeted monitoring and management in Romanian coastal waters and demonstrates a transferable three-axis framework—terrigenous, carbonate, and anthropogenic—for separating matrix effects from human pressures on river-modulated shelves, thereby strengthening the baseline definition, source attribution, and trend detection.

Implications for Monitoring and Management

The baselines and multivariate structure established in this study provide a process-based reference for ecological status classification and trend detection on the Eastern Romanian shelf. In practical terms, the results support the targeted surveillance of nearshore hotspots, particularly in sectors influenced by ports, urban discharges, and intense seasonal activity, where localized enrichments can occur despite an otherwise natural background, texture- and carrier-aware interpretation of metals using grain-size and/or Fe (or Al) normalization to distinguish lithogenic control from genuine contaminant inputs and to avoid misclassification driven by sediment sorting and mineralogical variability, and integrated contaminant and temporal diagnostics, combining organic tracers (e.g., ΣPAHs and, where feasible, black carbon) with radiometric constraints (e.g., excess ²¹⁰Pb for sediment accumulation and focusing) to resolve time-dependent changes and improve source apportionment.

Overall, the Eastern Romanian shelf investigated here is characterized by predominantly natural (lithogenic) regulation of trace-metal distributions, low-to-moderate organic contamination with a mainly petrogenic signature, and spatially restricted pyrogenic and metal hotspots in proximity to human activity. The multi-proxy geochemical–radiometric framework strengthens management relevance by explicitly separating lithofacies-driven background variability from localized anthropogenic signals, thereby establishing a defensible baseline for future assessments under evolving climatic forcing, shoreline development, and changing socio-economic pressures. Radiometric screening also adds value by documenting both natural-series radionuclides and anthropogenic markers (e.g., ¹³⁷Cs) in the sediment archive, while neutron-activation observations provide supplementary support for the broader elemental and radiological characterization.