Introduction

The non-indigenous benthic bivalve Ruditapes philippinarum (Adams & Reeve, 1850), commonly named as Manila clam, is a well-succeeded invasive species in worldwide estuaries and coastal systems. It represents one of the most commonly consumed bivalve species extensively cultivated all over the world [1]. Originally native to the Indo-Pacific region, it was introduced into the west coast of North America, the East Atlantic (Portugal, France, Spain, Ireland, England) and Mediterranean European coasts (France, Italy) [2]. In Portugal, it was introduced in the 1980 in Ria Formosa (South of Portugal), rapidly invading estuaries and coastal systems over all of the country, including the Sado estuary [3] and the Tagus estuary [2, 3]. With a large geographical spatial distribution and increasing abundance in recent years, R. philippinarum gained a significant economic relevance for fisheries and aquaculture, being intensively harvested in Europe [4,5,6,7]. In Portugal, it is one of the most important commercial bivalves [1], with an intensive exploitation, particularly in the Tagus estuary, though it is an illegal harvesting, involving a serious risk for human consumption due to microbiological and metal contamination.

The Manila clam has also been recognized as an appropriate bioindicator and/or sentinel species of environmental quality in aquatic systems [8, 9]. Filter-feeding species have a natural bias to absorb and accumulate metals, resulting on multi-element signature profiles, which can be used as a traceability tool to identify the habitat origin and to assess the ecological status in coastal and estuarine systems [10,11,12]. High ability of Manila clam to accumulate elements in its soft tissues and shells has supported its use as a bioindicator of pollution for specific elements in estuarine environments, e.g. Pb, As, Hg and Cd [13,14,15]. High filtration rates and the large feeding spectrum of this bivalve [16] allow the accumulation of large number of metals aggregated in a suspended particles, revealing a higher predisposition to uptake metals and metalloids than other bivalve species [4, 8]. Nevertheless, the bioaccumulation behaviour is influenced by the bivalve’ intrinsic species features, such as sex, age and physiological conditions (e.g. the reproduction physiology), and by the interactions with the abiotic conditions, particularly within the period of chemical exposure [8, 17,18,19]. The bivalve multi-element signature profiles are deeply dependent on the estuarine environmental conditions, namely, the salinity gradient, pH, temperature and tidal variations and turbidity [20,21,22,23,24].

The chemical profile that successfully discriminate specimens from different sampling sites is based on the link between species and their environmental conditions: trace elements such as Fe, Cu, Zn, Mn, Mg and Cr are related with bivalves’ growing and cell metabolism [18, 25, 26]; the presence of xenobiotic elements as Hg, Pb, Cd and As are related with the toxicity levels and the tolerance in the specimens that are influenced by anthropogenic disturbances [7, 27, 28]; Ni, Cu and Co are mainly related to the estuary lithological nature, becoming important to determine the organisms’ geographic origin [29]; and the trace elements such as Sr, Sc, Se, Cs, Ce, Eu and Th are geologically related to sediment morphogenesis [29, 30]. Bivalve chemical profiles are also related to their life cycle, reflecting different clams’ physiological conditions for bioaccumulation [31]. In the Tagus estuary, the reproductive cycle showed a synchronous gender gonadal development occurred in April and May, followed by an extensive spawning until December [32].

The ICP-MS (inductively coupled plasma mass spectrometry) is a highly sensitive analytical technique, and its high accuracy, precision and sensitivity allows the multi-elemental measurement at trace levels (ranging from part per trillion (ppt) to part per million (ppm)) as a multi-element screening [33, 34]. This analytical tool allows to distinguish different element species and isotopes in a single analysis, being effective to determine the specimens’ multi-element composition profiles with a high cost-effectiveness [25, 35, 36]. This technique performed in bivalve soft tissues revealed to be useful to identify the bioaccumulation throughout the multi-element composition of populations under metal pollution disturbances [7, 37], generally reflecting local environmental conditions [26]. The applicability of this technique in soft tissues could be explored to determine the bivalves’ origin by using trace elements as site-specific descriptors. The multi-element tag to discriminate the spatial distribution patterns of the non-indigenous R. philippinarum populations could be detected using the ICP-MS technique, a high sample throughput tool for geographic origin determination [38].

The main aim of this study was to analyse the spatial and temporal distribution of the multi-element signatures of the Ruditapes philippinarum populations collected in the Tagus and Sado estuaries, located in the SW coast of Portugal, to assess the trace elements as a traceability tool to support management decisions of the human activities in coastal ecosystems. The following research question was addressed: Are there significant differences on the R. philippinarum multi-element signatures between populations of the Tagus and Sado estuaries, within sampling sites of each estuary and between sampling occasions? Based on expected differences of the environmental conditions in both estuaries and the differences in bivalves’ life cycle, we hypothesized to find significant differences in clam’s chemical profile between both estuaries, between sampling sites within each estuary and between sampling occasions. We hypothesize that differences between estuaries and sampling sites should mainly be explained by chemical elements related with geomorphological and anthropogenic inputs, while the temporal differences should arise from the biogenic processes of the bivalves.

Material and Methods

Study Areas

Sampling sites were located in intertidal mudflats of the Tagus and Sado estuaries (southwest coast of Portugal). Both estuaries are geologically related to the Iberian Pyrite Belt and earlier mining activities, which explain the high concentration of the elements Fe, Cu, Ni, Co and rare earth elements (REE). Despite the proximity of the estuaries and the common hydrologic basins, there are differences in the distribution patterns of those elements in sediments among estuaries [29, 39, 40].

The Tagus estuary (38° 44′ N; 09° 08′ W) is one of the largest estuarine systems in Europe, with approximately 320 km2 and is located in the most populated area of Portugal. Nearly 40% of this area consists of intertidal flats and saltmarshes, which are partly included in the protected area designated as “Tagus Estuary Natural Reserve” [41]. The Tagus estuary is a semi-diurnal mesotidal system, with a tidal amplitude varying between 4 and 1 m during spring and neap tides, respectively [42]. The salinity gradient is strongly influenced by the Tagus river flows (mean value of 400 m3 s−1) changing with seasonal and inter-annual conditions [43, 44]. Water temperature ranges from 8 to 26°C [41]. The lower Tagus estuary southern branch has small enclosed bays (e.g. Montijo, Barreiro and Seixal bays) with wide intertidal areas, including important saltmarsh ecosystems, characterized by a high water residence time (ranging from 6 to 65 days) but strongly influenced by tides. This large estuary has a high socio-economic importance, supporting one the major Portuguese international harbours and several chemical, petrochemical, metallurgic, shipbuilding and cement manufacture industries [9]. Industrial and urban pollution sources have been mitigated through the construction of sewage treatment plants [22]. Despite intense anthropogenic disturbances observed in this estuarine ecosystem, it provides suitable environmental conditions to support a high abundance of R. philippinarum, which is currently a dominant bivalve species in the Tagus estuary [45]. The economic importance of the Manila clam harvesting at the Tagus estuary combined with high microbiological contamination levels triggered public health hazards related to the bivalve consumption [5].

The Sado estuary (38° 31′ 14″ N, 8° 53′ 32″ W) is the second largest estuarine system in Portugal, with an area of approximately 240 km2 and, similar to the Tagus estuary, with a high socio-economic importance, supporting important industrial and harbour-associated activities [46], with copper mines; heavy industries such as a paper mill, shipyards, pesticide and fertilizer factories; thermoelectric plant and intense aquaculture production [46, 47]. The upstream areas of the Sado estuary are intensively explored with rice fields, which are responsible for high inputs of fertilizers and pesticides discharged in the river [48]. Most of the estuarine area is classified as a protected area, designated as “National Reserve of the Sado estuary”. The Sado estuary has a semi-diurnal mesotidal system with tidal amplitude varying between 1.6 and 0.6 m during spring and neap tides, respectively. Salinity is influenced by the Sado river flow (annual mean of 40m3.s−1) changing with seasonal and inter-annual conditions [49] and with temperature ranging from 10 to 26°C. This system is partially separated by intertidal sandbanks (Troia beach) and linked to the ocean by a 50-m deep channel [49]. The intertidal area has approximately 78 km2 [50], and 30% of this area consists of salt marshes and intertidal flats and the water residence time of the estuary being approximately 21 days. Bivalves’ harvesting activity is residual when compared to the Tagus estuary. R. philippinarum abundance is much lower, most likely because the Sado estuary might be in an early expansion progress for this invasive species [45].

Sampling Design

The selection of the sampling sites to compare spatial and temporal distribution patterns of clam chemical profile signatures was based on the following combined criteria: (i) the occurrence of R. philippinarum populations in the Tagus and Sado estuaries [2, 5, 51] and (ii) the gradient of hydrological and geomorphological characteristics of both estuaries. Accordingly, three sampling sites were selected in the Tagus estuary: Alcochete (AL) located at the estuary uppermost section, next to the “Tagus Estuary Natural Reserve”, and strongly influenced by the Tagus river flow; the Montijo bay (MT), located in the surroundings of heavy industrialized areas; and the Seixal bay (SE), characterized by the intertidal important salt marshes [52] (Fig. 1a). Three sampling sites were selected in the Sado Estuary: Gâmbia (GA), located at the northern branch of the estuary, with an important oyster nursery production and a large paper mill and shipyards in its vicinity; Herdade do Pinheiro (HP), located in the northern branch of the estuary within the “Sado Estuary Natural Reserve”, is highly influenced by the Sado river flow and classified as a Special Protection Area for the conservation of wild birds; and Cabeço do Ratão (CR), located in a small sand bank at the southern branch of the estuary, next to Tróia peninsula (Fig. 1b). The sampling design and GPS locations are summarized in Table 1.

Fig. 1
figure 1

Location of the sampling sites of R. philippinarum and associated sediments. a Tagus estuary: Alcochete (AL), Montijo (MT) and Seixal (SE). b Sado estuary: Gâmbia (GA); Herdade do Pinheiro (HP) and Cabeço do Ratão (CR)

Full size image
Table 1 Sampling design: estuaries, sites and sampling occasions. GPS locations and sample code are also indicated
Full size table

Environmental Data

Temperature (°C) and salinity of the overlying water above the sediment were measured in situ, at each sampling site and sampling occasion to assess the environmental conditions of the collection sites (Supplementary material - Table S1). Data of the sediment grain size and organic matter content (OM) in the sediments is also included in the Supplementary information (Table S2).

Biological Data

R. philippinarum specimens were randomly collected at each sampling site during spring low tide at three sampling occasions: (1) May 2018 and May 2019, corresponding to the reproduction period for this species, and (2) January 2019, which is representative of the resting phase for this species in Portuguese estuaries [32]. Thirty adult specimens were collected at each sampling site and sampling occasion. Collected specimens length varied between 3.5 and 4.5 cm, representing adults with sexual maturity [32]. Sediments associated with the clam specimens were sampled (approx. 5 cm deep, which represents the bivalves living area) at each site and sampling occasion, to relate the multi-element signature of the clams’ soft tissue with that of the sediments. The samples were transported on ice to the laboratory and stored at −20°C.

Sample Preparation for ICP-MS Analysis

Due to the different matrix and complexity of the studied samples, two independent procedures for sample preparation and digestion were established in this work, in order to determine the concentration of 18 trace and major elements (Sc, Sr, Se, Cs, Ce, Ni, Sr, Co, Cu, Cr, Mn, Fe, Zn, As, Cd, Pb, Eu and Th).

Bivalves Sample Preparation and Digestion

A set of 10–15 clams were used to minimize the chemical variability of the clams and to ensure the reproducibility of the data of each clams site sampling. Specimens were rinsed with Milli-Q water (18.2 MΩ.cm) in order to avoid any contamination due the presence of other species like algae or other epibionts and sediments. After the pre-treatment, the bivalves’ soft tissues were removed from the shell with a plastic knife, rinsed between 2 and 3 times with Milli-Q water (18.2 MΩ.cm) and stored overnight at −20°C. All samples were freeze-dried at −80°C for approximately 72h, after which they were milled to homogenize the sample and to enhance the digestion performance.

The acidic digestion procedure of the clams was based on the method previously described in the literature [53], with some modifications. Briefly, approximately 100 mg of powdered clams, previously freeze-dried, were weighted in PFA Savillex® beakers and exposed to a pre-digestion step with 3 mL of HNO3 (65%, Suprapur, Merck®) at room temperature for 24h (note: this first step proved to be essential to allow for a full and safe digestion of the organic material). After the pre-digestion period, samples were digested in closed beakers on a hotplate at 125 °C for 3 h (during this time the digestion could be monitored through the formation of an orange atmosphere). Subsequently, all samples evaporated until dryness over a hotplate at 150°C, followed by addition of 2 mL of concentrated HNO3. The second digestion step was then started by heating up the samples at 125°C over 24h. The second digestion step was completed through addition, at room temperature, of 2 mL of H2O2 (30 %, OPTIMA grade, Fisher Chemicals). When the digestion of the organic matter was complete, the samples were allowed to dry on a hotplate and were re-suspended in 3 mL of Milli-Q H2O and 1.6 mL of concentrated HNO3, followed by dilution in PFA volumetric flasks with Milli-Q H2O up to 50 mL, yielding a final matrix of 2% HNO3 for analysis. Data reproducibility was ensured by means of digestion triplicates for all samples. A Certified Reference Material (CRM) of mussel tissue (SRM – 2976) was used and fully digested according to the described method, together with a blank control, for validation of the method.

Sediment Sample Preparation and Digestion

Prior to any sample treatment, sediments were placed in an oven and dried at 50°C, over 24h. Sediments were then sieved using two different mesh sizes (500μm and 1mm) in order to remove possible solid contaminants like shells and other associated debris. Finally, the sediments were milled using the Planetary Ball Mill PM 100 (RETSCH®) to yield a fine and homogeneous powder.

Sediments were digested under acidic hotplate attack, following the methods previously described in the literature (Makombe et al. 2017), with some adaptations. In PFA Savillex® beakers, sediments samples were weighted (≅ 100 mg), and a mixture of 0.5 mL HNO3 (65% v/v) (Suprapur, Merck®) and 2 mL of HF (50% v/v) (OPTIMA Grade®) was added. Afterwards, the closed beakers were placed over a hotplate at 125°C for 48h for a first cycle of digestion, after which the caps were removed and the solutions were allowed to concentrate almost until complete dryness (note: complete dryness will provoke the stabilization and precipitation on insoluble fluorides). A second acid attack was performed by adding 2 mL of aqua regia solution (freshly prepared) and left overnight at 150°C with closed vessels. After a second drying cycle was performed and 2 mL of concentrated nitric acid was added, the mixture was allowed to digest at 130°C, over more than 24h. The last digestion step was finished by adding, at room temperature, 1 mL of H2O2 (OPTIMA grade®), followed by a last drying cycle. The residue was then re-suspended in 3 mL of Milli-Q H2O and 1.6 mL of concentrated HNO3, followed by dilution in PFA volumetric flasks with Milli-Q H2O up to 50 mL, yielding a final matrix of 2% HNO3 ready for analysis. As mentioned before for the clam samples, data reproducibility was ensured by means of digestion triplicates for all the samples. A CRM of estuarine sediment BCR-667 was used and fully digested according to the described method, together with a blank control, for validation of the method. According to the concentration range of the analytes, the samples were diluted 10 or 100 fold when necessary.

ICP-MS Analysis: Operating Conditions and Analytical Parameters

Multi-elemental quantification was performed on an Agilent 8800 ICP Triple Quad (ICP-QQQ). Prior to the analysis, the equipment was calibrated with a tune solution from Agilent Technologies, and the sensitivity and resolution were optimized as well as the doubly charged ions (< 2.07%) and oxides species (< 1.04%) were minimized.

The external calibration method was used for the quantification, and a calibration curve with 8 levels (0, 50, 100, 200, 400, 800, 100 and 3000 ppb) was prepared using a multi-elemental solution from High Purity Standards (solution A, ICP-MS-68B). Along ICP-MS analysis, a solution containing three internal standards (101Ru, 103Rh, 193Ir) was added online to correct the data for possible drifts and matrix effects. Method precision was evaluated running each solution in triplicate reporting for each sample the relative standard deviation (RSD). The method was validated through the accuracy determination by means of analysis for the chosen CRM, every 10 samples measurements, which is presented and summarized in Table S3 and S4.

Experimental detection limits (LOD) were performed by measuring 11 replicates of a blank solution and 11 replicates of 400 ppb standard solution. Quantification limits (LOQ) were calculated from 10 times the detection limit obtained before (Table S3 and S4). The operational conditions are described in Table 2.

Table 2 ICP-MS operating conditions and instrument parameters
Full size table

Data Analyses

Multivariate ordination analyses were applied to identify the temporal and spatial changes of trace-elements concentrations (mg kg−1) in clams soft tissues and sediment samples, between “estuaries” and between “sites” under different physiological conditions, reproduction and resting stage, represented by “sampling occasions”. A principal component analysis (PCA) was performed using all multi-elements variables data. The elements were “grouped” based on (i) geomorphologic characteristics of each estuary (Se, Co, Ni and Cu), (ii) importance of clams metabolic processes (Mn, Fe, Zn and Cr) and (iii) elements derived from the anthropogenic inputs (As, Pb, and Cd). All elements (Sc, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Sr, Cd, Cs, Ce, Pb, Eu and Th) obtained from clams and sediment samples at different “estuaries”, “sites” and “sampling occasions” were expressed as mg kg−1. The PCA was applied separately for clam and sediment scaled datasets using the function “fviz pca_biplot” in the library “FactoMineR” using “R package” [54, 55]. To determine the contribution of each variable the square cosine (COS2) was used. The 3-factor permutational analysis of variance (PERMANOVA) add-on package [56] was applied to determine the significant differences between estuaries, sites and sampling occasions, using PRIMER v6 software package [57]. The design test included the following factors: “estuary” (2 levels, fixed), “sites” (6 level random and nested in estuary) and “sampling occasion”, (3 levels, fixed) for all variables analysed at p < 0.05 level. The analysis was performed for the 3 groups representing geomorphologic, metamorphic and anthropogenic elements concentrations resemblance matrix calculated based on Euclidean distances. Data was checked for a normal distribution and, when necessary, was log (X+1) transformed before analysis then normalized by subtracting the mean and dividing by the standard deviation for each variable. Redundancy analysis (RDA) was performed to answer how the geomorphological, metabolic and anthropogenic clam’s profile can be explained by the concomitant geomorphological, metabolic and anthropogenic sediment characteristics. Response data consisted of three (geomorphological, metabolic and anthropogenic) normalized clam’s element matrices, and explanatory matrix consisted of the same set of geomorphological, metabolic and anthropogenic sediment’s element profiles. Variables of the explanatory matrix were log10 transformed. A forward selection procedure, using function ordiR2step() was used to select only significant variables (p<0.05). Variation inflation factors (VIF) was calculated to check for linear dependencies and to ensure that only variables with small VIFs (<10) were included. RDA analysis was performed in R [55] using “vegan” and “BiodiversityR” packages [58, 59].

Results

Spatial and Temporal Variability of Multi-elements in R. philippinarum

The mean ± SE of multi-element concentrations measured in R. philippinarum soft tissues are expressed in mg kg−1, DW (“dry weight”) and are summarized in Fig. 2 and Table S5 and S6. The concentrations of multi-elements analysed in clam’s soft tissues (Se, Co, Ni, Cu; Mn, Fe, Zn, Cr; As, Pb and Cd) were present in the clam’s soft tissues above the analytical method LOQ, allowing for the determination of the chemical signature distribution patterns in the Tagus and Sado estuaries.

Fig. 2
figure 2

Mean values ± SE, n=3 of trace and major elements (mg kg−1) registered in clam samples of the Tagus and Sado estuaries, at each sampling occasion. (a)Trace elements: geomorphologic, metabolic and anthropogenic tracers and (b) major elements: metabolic tracers in the Tagus and Sado estuaries in May 2018; c Trace elements and (d) major elements in Tagus and Sado estuaries in January 2019; (e) Trace elements and (f) Major elements in the Tagus and Sado estuaries in May 2019. All concentration values of major elements were log-transformed.

Full size image

The elements, aggregated by groups Cu, Co and Ni (geomorphologic); Mn (metabolic); and As and Pb (anthropogenic), registered high variability between estuaries and sampling occasions (Fig. 2). In Tagus estuary, the highest concentration were observed in Alcochete and Montijo (TM18AL and TM18MT) in May 2018, whereas the lowest concentration were also registered in May 2019 in Seixal (TM19SE) and in January 2019 at Alcochete (TJ19AL) (see Table S5 and S6). In Sado estuary, the highest concentrations were observed in January 2019 and May 2019 at the H. do Pinheiro sampling site (SJ19HP and SM19HP), whilst the lowest concentrations were registered in Gâmbia (SJ19GA) in January 2019; the exception was observed for the element As, which showed the lowest values in May 2018 and January 2019 at C. do Ratão (SM18CR) (SJ19CR) (see Table S5 and S6). Throughout the study period, the Tagus estuary registered the highest concentrations of the elements and the highest variability among sites. However, it was in the Sado estuary that the highest concentrations of As and Cu were detected, respectively, in January 2019 and May 2018 (Fig. 2).

The PCA ordination analysis based on geomorphologic and anthropogenic elements explained 83.1% and 88.4% of the variability in the first two principal coordinates, respectively, clearly discriminating estuary origin of the clams’ population (Fig. 3 a, b, c). On the contrary the origin discrimination based on metabolic elements was less evident (Fig. 3b). The elements Fe, Co, Cr and Pb contribute to explain the variability in Tagus estuary, and the Cu, Cd, Fe and As presented the highest contribution in Sado estuary.

Fig. 3
figure 3

a Principal component analysis (PCA) biplot based on scaled geomorphologic element concentrations measured in clams, coloured by estuary “confidence” convex type. Variable’s vectors are presented based on their contributions to the principal components (gradient colours and transparency of vectors) with grey representing high contributions, yellow intermediate and blue representing very low contributions. The dots represent the cluster centroids for group variables. b Principal component analysis (PCA) biplot based on scaled metabolic element concentrations measured in clams, coloured by estuary “confidence” convex type. Variable’s vectors are presented based on their contributions to the principal components (gradient colours and transparency of vectors) with grey representing high contributions, yellow intermediate and blue representing very low contributions. The dots represent the cluster centroids for group variables. c Principal component analysis (PCA) biplot based on scaled anthropogenic element concentrations measured in clams, coloured by estuary “confidence” convex type. Variable’s vectors are presented based on their contributions to the principal components (gradient colours and transparency of vectors) with grey representing high contributions, yellow intermediate and blue representing very low contributions. The dots represent the cluster centroids for group variables

Full size image

The PCA ordination applied to metabolic and geomorphologic elements clearly discriminate the clams between sampling occasions (May 2018, January 2019 and May 2019), according to the life cycle :“May” is the reproductive period and “January” is the rested period. The elements Fe, Cr, Cu and Co present the highest contribution to explain the temporal variability (Fig. 4a, 4b, 4c).

Fig. 4
figure 4

a Principal component analysis (PCA) biplot based on scaled geomorphologic element concentrations measured in clams, coloured by sampling occasions “confidence” convex type. Variable’s vectors are presented based on their contributions to the principal components (gradient colours and transparency of vectors) with grey representing high contributions, yellow intermediate and blue representing very low contributions. The dots represent the cluster centroids for group variables. b Principal component analysis (PCA) biplot based on scaled metabolism element concentrations measured in clams, coloured by sampling occasions “confidence” convex type. Variable’s vectors are presented based on their contributions to the principal components (gradient colours and transparency of vectors) with grey representing high contributions, yellow intermediate and blue representing very low contributions. The dots represent the cluster centroids for group variables. c Principal component analysis (PCA) biplot based on scaled anthropogenic element concentrations measured in clams, coloured by sampling occasions “confidence” convex type. Variable’s vectors are presented based on their contributions to the principal components (gradient colours and transparency of vectors) with grey representing high contributions, yellow intermediate and blue representing very low contributions. The dots represent the cluster centroids for group variables.

Full size image

The PERMANOVA analysis, based on metabolic elements, showed significant differences between “estuaries”, “sites” and “sampling occasions” (p≤ 0.05), and significant interaction (p≤ 0.05) was detected between the factors “sites” × “sampling occasions”. Based on geomorphologic elements, significant differences were obtained between “sites” and “sampling occasions” (p≤ 0.05), and the anthropogenic elements showed significant differences only between “sites” and significant interaction between the 3 factors “estuary” × “sites” × “sampling occasions” (p≤ 0.05) (Table 3). Individual pairwise comparisons on the interaction factor “sites (estuary)” showed higher variability of the geomorphological and metabolic elements between sampling sites along the sampling occasions in Tagus estuary (p<0.001). The significant differences between sampling occasions were obtained between clams collected in January 2019 from those collected at May 2018 and 2019, (p<0.001). Individual pairwise comparisons on the interaction factor “sites (estuary)” based on anthropogenic elements showed significant differences between clams from all sites in each estuary (p< 0.05).

Table 3 Details of three factor PERMANOVA test with “estuary” (“Tagus” and “Sado” two level, fixed), “site” (“Alcochete”, “Montijo”, “Seixal”, “Gâmbia”, “H. Pinheiro” and “C. Ratão” six level, random) and “sampling occasion” (“May 2018”, “January 2019” and “May 2019” three level, fixed)
Full size table

Spatial and Temporal Variability of Multi-elements in Estuarine Sediments

The mean ± SE of multi-element concentrations measured in the estuarine sediments collected simultaneously with clams, at the sampling sites and at different sampling occasions, are expressed in mg kg-1, DW “dry weight” and are summarized in Fig. 5 and Tables S7 and S6.

Fig. 5
figure 5

Mean ± SE, n=3 of trace and major elements concentration values (mg kg−1) in sediments collected at sampling sites in the Tagus and Sado estuaries, at each sampling occasion. (1) Trace elements geomorphologic tracers and (1b) major elements metabolic tracers in the Tagus and Sado estuaries in May 2018; (2) trace elements geomorphologic (2b) major elements metabolic tracers in the Tagus and Sado estuaries in January 2019; and (3) trace elements geomorphologic and (3b) major elements metabolic tracers in the Tagus and Sado estuaries in May 2019. All concentration values of major elements were log-transformed

Full size image

The concentrations of 14 elements analysed in the sediment sampled (Cr, Co, Ni, Cu, Cd, Cs, Ce, Pb, Eu, Th, Sc, Fe, Mn and Zn) were above the analytical method LOQ, allowing the determination of the chemical signature distribution patterns in the Tagus and Sado estuaries. The elements were grouped according to the geomorphologic origin (e.g. Sc, Sr, Se, Cs, Ce, Ni, Sr, Co, Cu, Eu and Th) and metabolic function in bivalves’ life cycle (e.g. Zn, Mn and Fe). In both estuaries, sediment multi-element concentrations registered low variability between sampling occasions. Several geomorphologic tracers such as Se, Cu and Ni revealed high variability between estuaries and sampling occasions (Fig. 5). Nevertheless, higher element concentrations in the sediments were detected in Tagus estuary than in Sado estuary.

The first two axes of the PCA ordinations accounted for 90.7 % of data variability, highlighting the discrimination between the sediment element concentrations of the upstream and downstream sampling sites (Fig. 6 a and b).

Fig. 6
figure 6

a Principal component analysis (PCA) biplot based on scaled geomorphologic elements concentrations measured in sediments, coloured by estuaries with “confidence” ellipse type. Variable’s vectors are presented by their contributions to the principal components (gradient colours and transparency of vectors) with grey representing high contributions, yellow intermediate and blue representing very low contributions. The dots represent the cluster centroids for group variables. b Principal component analysis (PCA) biplot based on scaled geomorphologic elements concentrations measured in sediments, coloured by sampling occasions with “confidence” ellipse type. Variable’s vectors are presented by their contributions to the principal components (gradient colours and transparency of vectors) with grey representing high contributions, yellow intermediate and blue representing very low contributions. The dots represent the cluster centroids for group variables

Full size image

The PERMANOVA analysis detected significant differences of the geomorphologic elements between “sites” (p≤ 0.05) and significant interaction (p≤ 0.05) between the factors “sites” × “sampling occasions”. In accordance with PCA results (Fig. 6a, b), individual pairwise comparisons on the interaction factor “sites (estuary)” × “sampling occasion” showed significant differences in geomorphologic elements, between sampling sites and throughout sampling occasions (p<0.001). No significant differences were observed across sampling occasions at each site (May 2018, January 2019 and May 2019).

RDA Analysis

RDA analysis was significant only for geomorphological-based dataset and indicated that the spatial distribution of bivalve’s chemical signature was mainly driven by the concentration of Ni in Tagus estuary and Cu in Sado estuary. Sediment’s elements, Cs, Sc and Ce, explained a significant variation in clam’s chemical profile (p=0.001; Adj R2 = 0.38 and F= 12.01) responsible for the spatial differentiation between Tagus and Sado estuaries (Fig. 7).

Fig. 7
figure 7

Redundancy analysis (RDA) biplot based on scaled geomorphologic elements concentrations measured in sediments and clams, coloured by estuaries with “confidence” convex hull type. Variable’s vectors are presented by their contributions. Only significant vectors are displayed. On red there are displayed clam’s elemental scores.

Full size image

Discussion

Most estuaries worldwide are anthropogenically disturbed as they are considered filters of neighbouring contaminants before entering the ocean. Most of these contaminants are toxic, persistent and bio accumulative [60]. The bivalves have a natural tendency to accumulate these chemical elements, creating multi-element signatures, which reflect their habitat conditions [61]. The main aim of this study was to compare elemental distribution patterns of clams’ populations within and among estuaries and between sampling occasions and to identify the most important chemical elements that contribute to the observed differences. According to previous studies, we expected to observe differences in clams’ chemical profiles even in spatially close populations [25, 36].

This study was done to investigate if the multi-element analysis in clam soft tissues can be used in the future as a tool to distinguish clams from two neighbouring estuaries, sampling sites and from different reproductive stages. Since the elements assimilation in clam soft tissues is strongly influenced by several physical-chemical factors [11, 13], the studied elements were grouped by their main functional attributes, revealing distinct perspectives of the geomorphologic, metabolic and anthropogenic influences in the biological systems.

Spatial Elemental Distribution Between Estuaries and Sites

Metabolic element concentrations in clams’ soft tissues showed significant differences between estuaries, sites and sampling occasions, and these differences were likely explained by the dynamics of the life cycle related metabolic processes. The clams’ assimilation mechanism determines the fate and the potential effects of elements in their tissues; e.g. Mn, Fe, Cu and Zn are essential in the bivalves metabolic functions and energy management [18, 25, 26].

The concentration of anthropogenic elements in clams’ chemical signatures did not display significant differences between estuaries but demonstrated significant differences between sites, which could be related to the influence of local unspecific pollution sources in each estuary. The presence of xenobiotic elements (e.g. Pb, Cd and As) in clam’s soft tissue are the result of the efficiency of the tolerance mechanisms [18, 19, 62].

Geomorphologic elements did not display significant differences neither between estuaries nor between sites, likely due to similar lithologic origin of both estuaries [29, 50] that represent the same hydrological basin named “Tagus-Sado basin” (SNIRH) [63].

In both estuaries, the highest elemental concentration values were obtained in the clams collected in salt marshes and intertidal flats areas (Tagus estuary: Alcochete (AL), Montijo bay (MT) and Sado estuary: H. do Pinheiro (HP)), characterized by the high water residence time, organic matter content and prevalence of fine sediment, that all together favour the sulphide minerals aggregation to sediment suspended particles [64].

The elements As, Cd, Cu, Cr and Fe explained between sites spatial variability of the clam multi-element profiles in Sado estuary where the highest concentration of As and Cd was registered in the clams collected at sites in the proximity of pollution sources associated to the neighbouring activities, e.g. pesticides from intensive agriculture in rice fields, shipping and aquaculture activities [45, 48, 65]. Sado estuary is also influenced by the Iberian Pyrite Belt, and earlier mining activities can account for the high abundance of Cu, Cr and Fe [29, 39, 40].

The concentrations of Co, Zn, Ni and Pb in clams’ soft tissues explained the sites spatial variability of Tagus Estuary. Zinc is the most abundant element, and it was previously observed in high concentrations of salt marshes in the Tagus estuary [66], most likely due to intensification of agriculture and industrial activities [22].

Temporal Elemental Distribution

Multi-element composition of the clam soft tissues sampled at different sampling sites in both estuaries in May 2018, May 2019 and January 2019 showed an evident temporal pattern. The highest concentrations of elements related with metabolic processes (Mn, Zn, Fe and Cr) and geomorphologic characteristics (Cu, Se, Co) were measured during the reproduction periods (May 2018 and May 2019), which coincide with higher clam’s assimilation efficiency and physiologic requirements. The temperature is the main factor for gametogenesis maturity [32], which could explain differences in chemical signatures between sampling occasions (Table S1), reflecting different clams’ physiological conditions for bioaccumulation. For instance, Fattorini et al. (2008) revealed the importance of inter-annual shifts in the gametogenesis as a response of trace metal variations in mussels from Adriatic Sea. These inter-annual variations are due to changes in water temperature, salinity and precipitation conditions that highly interact with organisms physiological conditions [13, 62, 67]. A similar temporal pattern of the clam chemical signatures was observed by Zhao et al. (2013), which was described as a response to shifting environmental conditions and clams’ different element assimilation.

The high variability in clams’ chemical signatures between sampling occasions May 2018 and January 2019 was mainly associated to Se, Co, Ni and Mn concentrations. These elements are known to be highly influenced by the seasonal and inter-annual variations promoted by the freshwater inputs, precipitation and anthropogenic sources [22, 65], whilst there were significant differences in clams’ elemental composition patterns between sampling occasions. Spatial patterns were more evident and they are likely associated with different environmental conditions in each site.

Multi-element Profiles—Sediment vs Clams

Sediment properties such as sediment grain size, organic matter and chemical contents indirectly determine the clam chemical signatures [6, 13, 21, 68]. With the results obtained in clams and sediment chemical profiles, it was possible to establish a similar distribution pattern only between sites within estuaries and across sampling occasions. However, with these results, we would expect that the sediment’s chemical profile would be highly related with the clams’ signature profile, allowing the identification of the populations’ origin. Meanwhile, it was solely possible to establish a direct link between some specific geomorphological elements of sediment and clams in both estuaries, being Cu and Ni the potential drivers of the origin of clams’ populations and Cs, Sc and Ce being responsible for distinct sediment geomorphologic profiles between estuaries. As most of all rare earth element (REE) [69], these elements showed low temporal and spatial variability among sites revealing to be a good geographical tracer to differentiate estuaries, although these elements are difficult to detect in a measurable level in clam soft tissues, being considered non-essential elements for the physiologic requirements of the specimens.

Conclusions

The spatial and temporal distribution of the multi-element signatures of the non-indigenous bivalve Ruditapes philippinarum of the Sado and Tagus estuaries were analysed considering different categories of the chemical elements: estuarine geomorphology sources (Se, Co, Ni and Cu), elements with function in metabolic processes of the clams (Mn, Fe, Zn and Cr) and elements derived from the anthropogenic inputs (As, Pb and Cd). This approach allowed to explore the contribution of each elemental category as a tool to identify the habitat conditions and the geographical origin of the clams. The elements that most contributed for the spatial and temporal distribution of chemical signatures of the Tagus estuary clam populations were Zn, Co, Ni and Pb, whilst for Sado estuary were Cu, Fe, Cr, As and Cd. These elements, representative of all elemental categories (geomorphologic, metabolic and anthropogenic), showed a high spatial variability and a low temporal variability with some elements directly linked to the local anthropogenic pressures which can be useful to obtain local tracers.

The multi-element signatures of R. philippinarum as a natural tag derived from the physical and chemical conditions of its habitat. The ICP-MS technique applied to an ecological perspective, it is an important approach to develop a potential rapid tool to use as ecological monitoring and habitat assessment, including for spatial and temporal habitat discrimination of estuarine populations.