A geo-chemo-mechanical study of a highly polluted marine system (Taranto, Italy) for the enhancement of the conceptual site model

The Litho-technical characterization of the deposit

The litho-technical characterisation of the sediments has resulted from: the geological inspection of the cores in the liners and of the undisturbed geotechnical samples; the paleogeographic reconstruction of the soil deposition^29,39; the soil geotechnical index properties; the geochemical and the mineralogical analyses. Here-forth, Fig. 7a reports the litho-technical section N–N′ whose trace is shown in Fig. 7b.

A First litho-technical unit, hereafter 1^stLTU (light yellow colour in Fig. 7a), of about 1.5 m thickness, has been found to cover the whole deposit. It is formed of either clay with silt, or sandy to slightly sandy silt with clay, deposited in recent times up to present, according to the sedimentology and paleogeographic studies. The corresponding grading curves (Fig. 8) show that its clay fraction, CF, varies in the range 27–53%, its silt fraction, MF, in the range 39–57%, and its sand fraction, SF, is minor, except for site S1, close to the Porta Napoli channel (Fig. 2). It is rich in organic matter and the pocket penetrometer S_u data (S_u < 20 kPa) prove its largely fluid consistency (Table 1). The coefficient of permeability, K, measured by oedometer testing on the samples taken by scuba divers within this unit (M boreholes), varies in the range 10^–8–5 × 10^–9 m/s.

Given the very low consistency of the 1^stLTU sediments, it is very likely that these have been either remoulded, or resuspended all way through their history, due to the navigation activities in the south of the I Bay and to the dragging of ship anchors (Figs. 1 and 2). Furthermore, Mastronuzzi et al.⁷ provided evidence of the occurrence of important flooding events in the last two centuries (1883, 1996 and 2005), causing resuspension and redeposition of the 1st LTU sediments. Given so, the exact age of the sediments within this unit cannot be assessed.

The underlying Second litho-technical unit, hereafter 2nd LTU (light orange colour in Fig. 7a), is on average 6 m thick and it is formed of grey-coloured sandy or clayey-sandy silt, or clay with silt, of consistency varying from fluid to soft (S_u =  < 40 kPa) and permeability, K, on average about 10^–9 m/s, if a sand layer occurring in the western part of the N–N′ section is excluded (Fig. 7a). Accordingly, the grading curves of this unit show that SF ranges between 3 and 24%, whereas CF and MF vary in the intervals 22–39% and 43–59%, respectively (Fig. 8), except for the sand layer interbedded in this unit in the western part of section N–N′ (SF = 52.8–79.8%; Fig. 8). For this sand level, from medium-dense to loose, an average K = 10^–6 m/s has been measured by means of permeameter testing⁸⁵. According to both the paleo-geographic and the sedimentological analyses, also the 2nd LTU sediments, slightly coarser than the 1^stLTU sediments, are either present day or recent and derive from the erosion of calcarenites and parent coastal-alluvial, or marine formations in-land.

At larger depths, several boreholes cross a Third litho-technical unit, hereafter 3rd LTU (light to dark blue colours in Fig. 7a). This is formed of silt, clayey silt or silty clays (Fig. 8), including local sandier levels, gravel levels and peat. It is of lower permeability than the overlaying units (K = 10^–9–5 × 10^–10 m/s) and includes different sub-units. The lower portion of the 3rd LTU is interpreted to be the result of a high energy fluvial deposition, most probably occurred during the Last Glacial Maximum (33–14 ka)⁸⁶. Above the fluvial sediments, the recognition of levels of peat with pulmonated gastropods suggests that part of this unit was deposited within a transient continental environment, most likely at the beginning of the Holocene (11–10 ka). This part of the 3rd LTU passes gradually to upper silts and clays containing lagoonal shells, which mark the beginning of the marine transgression, within a sheltered marine environment. About 9 ka years ago, this sheltered marine basin was affected by the fall of volcanic ashes, deriving from the Pomici di Mercato eruption of Vesuvius. As a result of such event, a whitish porous tephra layer, 5 to 40 cm thick, is found to occur locally within this unit, at about 19–21 m depth, as confirmed by the results of PXRF analyses discussed later.

The consistency of the 3rd LTU is minimum for the first very soft sub-unit (light blue in Fig. 7a; S_u < 20 kPa; Table 1), and higher at medium depth, in the soft second sub-unit (blue in Fig. 7a; S_u = 20–40 kPa; Table 1). Underneath, the sub-unit shown as dark blue in Fig. 7a is of firm consistency, i.e. S_u = 40–75 kPa, and overlays the deepest stiffest sub-unit, i.e. S_u = 75–150 kPa (Table 1), shown in cobalt blue. Boreholes S6 and S7 (Fig. 7a) do not cross the 3rd LTU, since the 2nd LTU overlies directly the ASP.

The ASP have been found at the bottom of all the boreholes in the section (Fig. 7a). They are grey-bluish silty clays (Fig. 8), from very stiff (S_u = 150–300 kPa, light grey, Table 1) to hard (S_u > 300 kPa, dark grey, Table 1), of low permeability (k = 10^–11 m/s laboratory measurement, k = 10^–10 m/s field measurements). Their top is deepest in the southern part of the I Bay and deepens to the west of the section (Fig. 7a)²⁹. The irregular surface of the ASP top represents the result of river erosion during the Last Glacial Maximum, before the deposition of the overlying units^29,39,87. Neither the GRA, or the CA (Figs. 3 and 4) were reached in the drilling operations along section N–N′; the top of CA reported in Fig. 7a (dashed green line) is that inferred by³⁰ through geophysical surveys.

The XRPD analyses, carried out on the S1 to S7 corings have shown that all the sediments have similar mineralogical composition to the ASP clays^{41,88,89,90,91,92}. As an example, Supplementary Fig. S4 illustrates the diffraction XRPD patterns for samples from borehole S2. The main mineralogical phases detected in the analyses are: clay minerals, quartz, carbonates (mainly calcite and aragonite), plagioclase and feldspar. Minor phases also occur, but they are not ubiquitous and are not homogeneously distributed within the sediments. The vertical profiles of mineral content in Fig. 9 testify an increase of clay mineral content with depth. This ranges from 24 to 40% in the 1st LTU, and from 40 to 60% in the 3rd LTU. In the 2nd LTU, the clay mineral content reaches an out of trend peak of 56% in a sample at 5 m depth b.s.f. in borehole S1, where an amorphous component is abundant and impacts the accuracy of the determination. The clay minerals are a mixture of illite, chlorite, kaolinite and inter-stratified illite–smectite phases (I-S); the smectite minerals are not detectable in the diffraction patterns as a single mineralogical phase.

The pore water salinity has been found to range between 30 g/L and 36 g/L within the 1^stLTU, and to be about 32 g/L within the deep sediments of the 3rd LTU⁵⁸. Therefore, the salinity of the pore water in th Holocene sediments remains high with increasing depth.

The limited variability in both the granulometry and the mineralogy of the sediments forming the different units, except for the coarser sediments locally interbedded in confined levels, is not consistent with the significant differences among the values of some geotechnical index properties recorded for the different units. This is the case for the values of the liquid limit, w_L⁹³, the plasticity index, PI (PI = (left( {{text{w}}0 – {text{wP}}} right)))⁹³ and the activity index, A = PI/CF, measured for the 1st LTU and 2nd LTU on one side, and those characterizing both the 3rd LTU and the ASP, on the other. Such differences are evident in Fig. 10, reporting the data in the Casagrande plasticity chart (Fig. 10a) and the activity chart (where PI is plotted versus the clay fraction CF, to characterize the activity index A; Fig. 10b). It is worth remarking that since w_L, PI and A are the geotechnical indices most closely related to the amount and mineralogy of the clay fraction, CF, and to the pore water salinity, their values should vary little among the 1st to the 3rd LTUs and the ASP. Conversely, the samples belonging to the 1st LTU are characterised by values of w_L (70–113%), PI (35–66%) and A (> 1.1) much higher than those recorded for the 3rd LTU and the ASP. Such anomaly is the first evidence of the dependence of the geotechnical properties of the sediments present in the I Bay on their content in contaminants and OM, as confirmed in the following.

Chemo-mechanical features

The integration of all the data acquired along section N–N′ (“The new multidisciplinary investigation” section; Fig. 6) is exemplified in Fig. 11. As such, the figure represents the geo-chemo-mechanical section N–N^′_GCM of use for the analysis of the spatial variability across the deposit of the lithological and geotechnical soil properties together with the contaminant concentrations. For each site the figure reports the litho-technical profile, aside the profiles of the soil granulometry and geotechnical properties: e₀, w_L and w_P, A, Gs, S_u. These profiles can be compared with those of the concentrations in: As, Cd, Cr, Cu, Hg, Ni, Pb, Zn, V, PCBs (expressed in term of sum of 31 PCB congeners), PAHs (expressed in terms of sum of 16 EPA congeners), TPH, and the OM content profile. All the data are plotted at the average depth of the sample they correspond to. Furthermore, for each contaminant profile both the threshold value set for the Taranto Site (yellow line) and that set by the National Environmental Law for industrial sites (red line) are reported. Such threshold values are listed in Supplementary Tab. S2.

For some of the geotechnical properties, Fig. 12 reports the profiles grouped together in a single plot, as well as a plot of all the OM content profiles. It is evident the ubiquitous presence in the section of a top layer, up to 2 m thick, which roughly corresponds to the 1st LTU, where the sediments are of highest liquid limit (w_L = 70–113%), plasticity index (PI = 35–66%) and activity index (A up to 2.5), and lowest soil specific gravity (G_s = 2.54–2.66), much lower than that typical for the clay minerals forming the skeleton of the sediments within this layer (e.g. G_s = 2.75–2.78 for montmorillonite, G_s = 2.74 for illite and G_s = 2.62–2.66 for kaolinite)⁹⁴. The low specific gravity is recognisably due, at least in part, to the major OM content in this layer, 15–18% (Fig. 14e), much higher than that usually measured in the sediments at the sea floor of open marine basins⁵⁶. In addition, the soils in this top layer have the highest void ratio, e₀ = 2.12–3.98 (Fig. 12f) and water content, w₀ = 72–157% (Fig. 12g), even higher than the corresponding liquid limit (w_L). Hence, their liquidity index LI:

is higher than 1 (Fig. 12h), as confirmed by the very low S_u values (Fig. 11). Therefore, the combined analysis of the data in Fig. 11 demonstrates that the very soft 1st LTU (S_u < 20 kPa), in the interface between the sediment deposit and the seawater column has geotechnical properties highly affected by the high OM content and, possibly, by the presence of the significant contaminant concentrations discussed in the following and leads to suppose that it is highly prone to remixing and resuspension.

Within the 1st LTU, only at site S1 the shallowest sample, which is of higher sand fraction (SF = 29.4%), has lower organic content (OM = 8–12% in Fig. 12e) than all of the other shallow samples in the section. This is likely to be due to the hydrodynamic conditions of the channel area where the site is located^95,96. In addition, the ratio of organic carbon to organic nitrogen, C_org/N_tot, measured in the I Bay in previous studies^24,97, reaches very high values close to both the sampling sites S1 and S2 (15 < C_org/N_tot < 65), suggesting that the organic matter in the channel area is largely allochthonous⁹⁸.

As to the physical–chemical properties, the sediments within the 1st LTU are characterised by neutral to slightly alkaline conditions, with pH values in the range 7.3–8.3 (Fig. 13b), and Eh in the range from − 400 to − 200 mV (Fig. 13c), indicative of a far more reducing environment with respect to that at the sea floor in both the Adriatic and the Ionian Sea, where positive Eh values are measured⁵⁶. The negative redox potential reveals a high rate of oxygen consumption in the I Bay, even before sediment deposition, which is likely to be due to aerobic microbial-mediated redox-processes, that can reduce the redox potential over − 300 mV and cause the total depletion of oxygen, used as terminal electron acceptor. After deposition, anaerobic mediated redox-processes are activated, such as anaerobic sulphate reduction (SO₄²⁻ to S²⁻) or anaerobic methanogenesis, this latter converting the CO₂ produced by the mineralisation of the organic matter, into CH₄.

The chemical data (Supplementary Table S1) plotted in Fig. 11 reveal that in the 1^stLTU, down to 1.5 m depth, some of the contaminants, either organic or inorganic, exceed either the red or the yellow thresholds cited above. In 4 of the 6 sampling sites along the section, S2, S3, S4 and S6, the concentration of metals Hg, As, Pb, Cu and Zn, either approach or exceed the Taranto Site (yellow) threshold (Fig. 11, Supplementary Table S2). In addition, values of Hg above the National Environmental Law (red) threshold are recorded at both sites S3 and S6, in the front of the Navy area. In particular, the highest concentrations of the cited metals measured in the samples taken in the first 0.5 m depth are: Hg = 15 ppm at S6, Pb = 262 ppm at S3, Cu = 88 ppm at S6, Zn = 403 ppm at S6, As = 45 ppm at S6, Cd = 1.16 ppm at S6 (Supplementary Tab. S1).

High metal concentrations are also recorded between 0.5 m and 1.5 m depth, although lower than in the top 0.5 m depth (Fig. 11). Unlike the other metals, the concentrations of Cr, Ni and V are well below the Taranto Site (yellow) threshold (Fig. 11) in the whole 1st LTU. The highest concentrations of As, Pb, Zn and Cu were recorded to occur within the top 0.5 m sediment layer of the 1st LTU also through the PRXF technique (Fig. 14). Since this technique detects the concentrations within spots of 6 mm diameter, very high concentrations of some metals were detected through its application also down to 4.0–5.0 m depth, as shown in Fig. 14.

Within the first 0.5 m stratum of the 1st LTU, the highest concentrations of the organic contaminants were recorded too (e.g. the overall sum of either the congeners of PCBs, or the congeners of PAHs), found to exceed even the National Environmental Law (red) threshold at several sites (Fig. 11). In particular, the highest values were recorded at S2 for ΣPAH_16EPA (9775 ppb), at S6 for ΣPCB₃₁ (6828 ppb), and at sites S2, S4, S6, S3 for TPH (750 ppm; Fig. 11). The concentration of the organic contaminants reduces below 0.5 m depth.

In the 2nd LTU and 3rd LTU, both the geotechnical indices PI and A reduce with increasing depth (Figs. 11 and 12), tending to the values typical for the Pleistocene ASP clays^33,41,99,100. Also the void ratio and the liquidity index decrease, due to both the compression of the sediments under burial and the reduction in content of either the OM or the contaminants.

However, in the 2nd LTU the soil void ratio, the liquidity index and he OM content of the fine sediments may be still quite high by 5 m depth (Figs. 11 and 12). As expected^97,101, a lower OM content is recorded in the sandy layer in the western part of the section (OM = 2.5%), which is, therefore, less capable to trap the contaminants, with respect to the surrounding finer soils. As to the physicochemical properties, in the 2nd LTU the pH values (Fig. 15b) change among the different sites, although they keep being from neutral and slightly alkaline in the eastern sites, S6 and S7, and tend to increase in the sandy layer intercepted in the western sites. The negative Eh values (Fig. 13c) of the 2nd LTU suggest that the process of degradation of the organic matter occurs under anaerobic conditions in both the first two units.

Contamination is heterogeneous within the 2nd LTU. A significant decrease in concentration of the metals with respect to the 1st LTU is detected, except for few local high values. For example, at about 2.25 m depth at sites S1, S2 and S3, Hg concentrations above the Taranto Site threshold (yellow) have been recorded. In addition, concentrations exceeding the Taranto Site threshold are recorded for: Zn at 2.5 m depth at site S3; Pb, down to 4 m depth at both sites S1 and S3; Cu, at 6 m depth at site S2. Furthermore, high concentrations of PAHs and PCBs have been recorded at sites S1 and S3, even exceeding the Taranto Site threshold (e.g. ΣPAH_16EPA ≅ 4592 ppb at 2.25 m b.s.f. in S1; ΣPCB₃₁ > 190 ppb at 3.55 m b.s.f. in S6). In addition, in S3, at 2.25 m b.s.f. the concentration of TPH (> 750 ppm) exceeds the National Environmental Law threshold.

The 3rd LTU, from 10 to 40 m b.s.f. (Figs. 7a and 11), is characterised by a low variability of the geotechnical index properties indicative of the soil composition (Fig. 11; average values: w_LAV = 50%, PI_AV = 28%, A_AV = 0.65 and G_sAV = 2.67), which are all very close to the values characterizing the ASP clays. The void ratio reduces with increasing depth (e₀ = 1.38–0.88), in a way compatible with the compression of the soil under burial (LI < 1; Fig. 12h). However, the OM content recorded in the soft to firm portion of the 3rd LTU (i.e. down to 35 m depth; light blue in Fig. 12e) is still quite high (OM = 6–12%), whereas it reduces to 5–6% only in the deepest firm portion (blue in Fig. 12e). Therefore, on the whole, the OM profile (Fig. 12e) suggests that the decomposition of the organic matter buried within the 3rd LTU has been anaerobic and so slow as to preserve high OM contents at depth long time after deposition. Such hypothesis is validated by the measured values of the redox potential and the pH in this unit, which are still typical of a reducing environment in large part of the unit (− 100 mV; Fig. 13), and tend to zero, or to positive values, only when the sediment becomes firm at large depth (blue in Fig. 13c). The high OM content of this unit justifies the G_s values slightly lower than those typical for the ASP clays (for ASP G_s = 2.73^33,41,99,100).

In the 3rd LTU, though, the concentrations of the metals and of the organic contaminants are very low at all sites (Fig. 13), with the only exceptions of the lithogenic metals Cr, Ni, Zn and V, whose concentrations keep being high in most part of the unit. At sites S1 and S2, both Cr and Ni even exceed the Taranto Site threshold, from 11 m b.s.f. and 17 m b.s.f. downwards, respectively (Fig. 11). Also the PRXF profiles (Fig. 14) indicate that the concentration of either Ni or Cr increases at the depth of transition between the 2nd LTU and the 3rd LTU. They also show an excess of As, Pb and Zn down boreholes S1, S2 and S3, at about 19–21 m depth, probably relating to the presence of the thin layer of volcanic soils (tephra) recognized through the geological analyses.

Finally, the samples belonging to the ASP formation (Figs. 11 and 12) are characterised by values of the index properties (i.e. w_LAV=48%, PI_AV = 24%, A_AV = 0.66, G_sAV = 2.71: w_0AV = 27%, e_AV = 0.74, LI_AV = 0.15) consistent with the average values usually measured for such clays either in land, or at depth below the Mar Grande seafloor^{32,33,41,99,100,102}. The OM content of the ASP clays is lower than that of the overlying units (OM = 6–8%), irrespective of depth (Fig. 12e) and the redox potential becomes positive (i.e. about 50 mV, Fig. 13c). As in the 3rd LTU, high concentrations of Cr, Ni, and V are recorded in the ASP clays, which demonstrate that the content of such metals recorded in both the 3rd LTU and the ASP clays is part of the skeleton of the sediments in these units, as further discussed in the following.

Source: Ecology - nature.com