Changes in mercury content in oysters in relation to sediment and seston content in the Colombian Caribbean lagoons

Study area

The Colombian Caribbean region is characterized by a bimodal climatic regime with a rainy and dry season influenced by the Intertropical Convergence Zone (ITCZ), generating periodic patterns (Restrepo & López, 2008). Trade winds predominate from December to April (dry season), changing direction to the southeast between April and November (rainy season) (Nystuen & Andrade, 1993).

The Ciénaga Grande of Santa Marta (CGSM) covers an area of 450 km² (González-Arteaga & Ricaurte-Villota, 2023), and was declared a Ramsar Wetland and Biosphere Reserve (UNESCO, 2001), Consiting of interconnected lagoons and a sandbar to the northeast separating it from the Caribbean Sea (Restrepo Martínez, 2004; Fig. 1A). The exchange of fresh and brackish water supports the development of Rhizophora mangle (red mangrove), providing substrate for the mangrove oyster (Crassostrea rhizophorae) (Rodríguez-Rodríguez et al., 2018). CGSM is a productive tropical ecosystem, yielding significant commercial catches of fish, crustaceans, and mollusks (Sánchez-Núñez, Viloria Maestre & Rueda, 2024).

Cispatá Bay (BhC), an estuarine system within the Sinú River delta, features fine to very fine sediments primarily influenced by the Sinú River. The 130 km² estuary is predominantly covered by mangroves (Castaño, Urrego & Bernal, 2010; Fig. 1A). Rainfall averages 66 mm in the dry season and 150 mm in the rainy season, with sediment discharge increasing from 3.1 kg/day to 11.5 kg/day during the rainy period (Rangel-Ch & Arellano, 2010). BhC salinity fluctuates seasonally due to rainfall, droughts, and the mixing of fresh and brackish water.

Field phase

Oyster samples were collected from three stations in CGSM and BhC, selecting sites that represented gradients of water inflow from the sea and freshwater sources that might carry contaminants. A single sampling was conducted during each period: the rainy season (November 2021) and the dry season (March 2022). At each oyster collection site, in situ measurements of temperature, salinity, pH, and dissolved oxygen were taken at a depth of 0.5 m using WTW 3110 and YSI Pro1030 multiparameter probes. Thus, for each ecosystem (CGSM and BhC) and climatic season (dry and rainy), three data points per physicochemical variable were obtained.

277 individual oysters were collected from CGSM, and 237 from BhC under the collection permit for wild species specimens of biological diversity for non-commercial scientific research purposes, granted by the Autoridad Nacional de Licencias Ambientales (ANLA) through Resolution 1271 of October 23th 2014, modified by resolutions 1715 of December 30th 2015 and 00213 of January 28th 2021, to the University of Bogota Jorge Tadeo Lozano (UTADEO). The samples were divided into six groups per station in each climatic season, with three groups consisting of juveniles (22.0–32.0 mm) and three groups of adults (35.0–56.5 mm) (Pacheco Urpí, Cabrera Peña & Zamora Madriz, 1983; Madrigal Castro et al., 1985). The result was a total of 71 composite samples across both ecosystems and climatic periods (Table S4).

At each station, specimens were primarily collected from the roots of mangrove trees. oysters were placed in plastic containers, were cleaned to remove any particles adhering to their shells (Fig. 1B), stored in pre-labeled, airtight polyethylene bag, and preserved in coolers with gel ice packs (∼4 °C).

To determine mercury content in seston, and to serve as food for filter-feeding organisms like bivalves, three water samples were collected at each station in 2.8 L amber flasks and kept cold (∼4 °C). After homogenization, samples were filtered through two Whatman GF/C glass fiber filters (47 mm diameter) per sample using a manual vacuum pump. Filters were then stored in hermetically sealed, pre-labeled polyethylene bags, dried in an oven at 45 °C for 24 h, and weighed on an analytical balance (Cogua, Campos-Campos & Duque, 2012).

In each station, three sediment samples were collected using a van Veen dredge. From each composite sample per station, 600 g of sediment was separated for mercury analysis, 75 g for organic matter determination, 75 g for redox potential measurement, and 300 g for granulometric analysis focused on the content of silt. Samples were stored in airtight polyethylene bags using a silicone scoop to avoid contamination, ensuring no contact with the dredge edges. Samples were kept chilled (∼4 °C) (Cogua, Campos-Campos & Duque, 2012).

Laboratory phase

To determine organic matter content, 5 g of dry sediment were placed in pre-weighed porcelain crucibles, subjected to calcination in a muffle furnace at 550 °C for 5 h, and then left in a desiccator for 2 h. Organic matter content was calculated based on the difference between the dry weight and the weight after calcination (Kenny & Sotheran, 2013).

For redox potential quantification, sediment samples were dried at 40 °C for 24 h. A portion of 25 g of sediment was then homogenized in 50 mL of deionized water using a VELP Scientifica magnetic stirrer for 30 min. Redox potential was measured with a YSI Pro1030 multiparameter probe equipped with an oxidation–reduction potential electrode, at a standard temperature of 25 °C (Aldridge & Ganf, 2003).

The 300 g of collected sediments were dried in an oven at 90 ± 5 °C. From the dried material, 100 g were weighed and it’s dispersed in one L of deionized water with 40 mL of 1% sodium hexametaphosphate ((NaPO₃)₆, 6.2 g/L) for 24 h to facilitate particle separation, following ISO 11277:2020 (ISO, 2020). Once dispersion was achieved, the samples were sieved covering the range from 1 mm to 63 µm. Although each sediment fraction was classified according to size (very coarse sand: 1 mm; coarse sand: 500 µm; medium sand: 250 µm; fine sand: 180–125 µm; very fine sand: 90–63 µm; silt: <63 µm), only the final weight of the <63 µm fraction (silt content) was recorded and used in this study for a descriptive analysis.

For the chemical analyses, all materials were pre-treated by purging with 5% nitric acid (HNO₃) and deionized water for 24 h to prevent contamination. Samples were handled with gloves, glass, or plastic materials to avoid contamination. Samples were then transferred to the Toxicology and Environmental Management Laboratory at the University of Córdoba, preserved at ∼4 °C, for mercury (Hg) quantification.

The anteroposterior length (APL) of the oyster was measured on the inner side of the ventral valve using a Vernier caliper (precision of 0.05 mm). soft tissues were removed and weighed using an analytical balance  ± 0.1 mg). For each sample, organisms of similar size were pooled, and soft tissues were placed in pre-weighed and labeled 30 mL glass vials. The vials were then lyophilized, and the final dry weight was recorded.

For mercury analysis in samples of seston, sediment, and oyster tissue underwent a pre-treatment and analysis at the Toxicology Laboratory, following a validated method. The pre-treatment included homogenizing the lyophilized samples (oyster tissue), weighing approximately 20–40 mg of each sample into nickel cells, and introducing them directly into a Direct Mercury Analyzer (DMA-80). This instrument operates based on atomic absorption spectrometry with thermal decomposition and gold amalgamation, in accordance with US Environmental Protection Agency (EPA) Method 7473 (USEPA, 2007).

During the combustion or pyrolysis step, all mercury compounds in the sample including volatile organic mercury species are thermally decomposed and converted to elemental mercury vapor (Hg°). Combustion by products, such as halogenated compounds and nitrogen and sulfur oxides, are removed by passing the carrier gas (oxygen) through a hot catalyst. The elemental mercury vapor, free from interferences, is transported by the carrier gas to a gold trap, where it is selectively adsorbed to form a gold–mercury amalgam (Milestone, 2020).

Once a sufficient amount of mercury has accumulated in the gold trap, it is rapidly heated to release the concentrated elemental mercury vapor. This vapor passes through an absorption cell in the path of an ultraviolet (UV) light beam at the mercury-specific wavelength (253.7 nm), allowing precise quantification.

Importantly, this closed system prevents any loss of mercury species including volatile organic forms ensuring accurate measurements. Therefore, there is no risk of mercury loss due to volatilization during the pre-treatment and analysis steps.

Sediment and seston samples were subjected to standard procedures for trace metal analysis in environmental matrices. The fraction smaller than 63 µm was selected, as this size fraction is known to better retain the largest amount metals. Both sediment and seston samples were digested using 5% nitric acid (HNO₃) to prepare them for total Hg determination. For seston, the laboratory adapted the digestion protocol originally developed for the sediment samples, ensuring methodological consistency between matrices. Total Hg concentration was subsequently quantified following EPA Method 7473 PLTX-017, which involves thermal decomposition, amalgamation, and atomic absorption spectrometry (EPA, 1998).

For analytical control, triplicate analyses of Hg solutions at different concentrations were performed. Calibration curves were established for three concentration ranges in all matrices: 0.005–0.02 µg Hg, 0.02–0.05 µg Hg, and 0.05–0.5 µg Hg. The coefficients of determination (R²) were 0.9993, 0.9986, and 0.999 for sediment; 0.9973, 0.9962, and 0.9979 for seston; and 0.9996, 0.9996, and 0.9989 for oyster tissue. Error percentages were kept below 15% for all three matrices. TORT-1 (lobster hepatopancreas) used as a reference material from the National Research Council of Canada (NRCC) for samples of oyster tissue, and GSD-10 for sediment and seston samples (Leng et al., 2013; Romero-Murillo, Campos-Campos & Orrego, 2023). Recovery percentages were 100 ± 1.4% for sediments (limit of detection (LOD) = 0.00073 µg/g Hg), 100 ± 5.4% for seston (LOD = 0.000015 µg/g Hg), and 100 ± 1.4% for oysters (LOD = 0.00073 µg/g Hg).

Data analysis

The bioconcentration factor (BCF) was calculated as the ratio of Hg concentration in oyster tissue to its presence in sediment (sd) and seston (st), expressed in parts per million (ppm, µg/g) in dry weight (d.w.). BCF was used to evaluate the efficiency of Hg accumulation in oyster soft tissue. Mountouris, Voutsas & Tassios (2002), BCF <1 suggests no metal accumulation, BCF ≥ 1 and <10 indicates accumulation and BCF ≥ 10 indicates hyperaccumulation of metal. The calculation was based on Mountouris, Voutsas & Tassios (2002) and Romero-Murillo, Campos-Campos & Orrego (2023). (1)${\displaystyle BC{F}_{sd}=\frac{{\left[Metal\right]}_{organism}}{{\left[Metal\right]}_{sediment}},BC{F}_{st}=\frac{{\left[Metal\right]}_{organism}}{{\left[Metal\right]}_{seston}}.}$ Permutation analysis of variance (PERMANOVA) was applied to compare Hg concentration in oyster tissue and its BCF between the two ecosystems (k = 2), the two climatic seasons (k = 2), the six stations (k = 6) and the two categorized size classes (k = 2). 9,999 permutations were performed using Euclidean distance and type III sum of squares. p-values were computed using Monte Carlo (MC) permutation testing only when unique permutations were less than 100 (Anderson, Gorley & Clarke, 2008).

Relationships between mercury (Hg) concentrations in sediment and seston were examined using Pearson’s and Spearman’s correlation analyses based on data distribution. Normality tests (Shapiro–Wilk) were performed on each dataset prior to analysis (Zar, 2010). Influences of environmental predictor variables on the oyster Hg concentration and Hg BCF in relation to seston were evaluated using a distance-based linear model (DistLM) with adjusted R² criterion and 9,999 permutations (Anderson, Gorley & Clarke, 2008).

To evaluate mercury contamination in bivalves, the Nemerow integral contamination index −P_c- was used (Ding et al., 2022). Calculations of P_c are based on the average value of the individual pollution index (P_avg), the maximum value (P_max) and the minimum value (P_min) of the collected data standardized to µg/kg wet weight (w.w.) of the bivalves. C_avg average concentration value, recorded in the data set evaluated, C_max and C_min are the maximum and minimum concentration values from the same data set, and S is the provisional tolerable weekly intake (PTWI) with Hg in adults (one µg/kg body weight per week) (FAO/WHO, 2010). The individual index values were calculated using the following formula: (2)${\displaystyle P_{avg}=\frac{C_{avg}}{S},P_{max}=\frac{C_{max}}{S},P_{min}=\frac{C_{min}}{S}.}$

Once the historical values of P_avg, P_max and P_min were obtained for each location by year, the calculation of P_cwas performed establishing (i) P_c ≤ 0.7 considered no risk, (ii) 0.7 <P_c ≤ 1 low risk, (iii) 1 <P_c ≤ 2 medium risk, (iv) 2 <P_c ≤ 3 high risk and (v) P_c >3 very high risk of contamination (Ding et al., 2022). ΣP_c is the sum of all P_c values divided by “n”, the total number of years evaluated per location in its historical record, ensuring that the values of P_max and P_min do not overestimate or underestimate the calculation of the contamination index for each metal evaluated, respectively. It was calculated using the following modification of the equation: (3)${\displaystyle \frac{\sum P_c}{n}=\sqrt{\frac{P_{avg}^2+P_{max}^2+P_{min}^2}{3}}.}$

A hierarchical clustering analysis was carried out using the squared Euclidean distance with Ward’s linkage, minimizing variability and producing uniformly sized clusters (Tudor et al., 2002; Tudor & Williams, 2004).

Environmental conditions in both coastal lagoons

In CGSM, the temperature during the rainy season was 31.17 ± 0.48 °C (n = 3), while in dry season it was 30.53 ± 0.84 °C (n = 3). In BhC the temperature during the rainy season was 28.97 ± 0.43 °C (n = 3), and in dry season it was 29.75 ± 0.03 °C (n = 3) (Fig. 2A, Table S1).

During the rainy season, CGSM had a higher pH value (8.77 ± 0.12, n = 3) compared to BhC (7.84 ± 0.09, n = 3), while during the dry season, were a decrease in pH in CGSM (8.44 ± 0.16, n = 3) and an increase in BhC (8.19 ± 0.003, n = 3) (Fig. 2A, Table S1).

The average dissolved oxygen content was higher in CGSM during both climatic seasons, with values of 7.83 ± 0.66 mg/L (n = 3) in rainy season and 7.39 ± 2.4 mg/L (n = 3) during dry season. In BhC, the contents were 4.32 ± 0.5 mg/L (n = 3) in rainy season and 5.52 ± 1.28 mg/L (n = 3) along the dry season (Fig. 2A, Table S1).

In CGSM, salinity values varied significantly between seasons, ranging from 2.47 ± 1.01 (n = 3) during the rainy season to 18.53 ± 6.33 (n = 3) in the dry season. In contrast, BhC showed less seasonal fluctuation, with average salinity levels from 24.9 ± 1.01 (n = 3) to 30.83 ± 0.56 (n = 3) across the two seasons (Fig. 2A, Table S1).

Organic matter in CGSM was also seasonally variable, with contents during the rainy season at 11.67 ± 3.27% (n = 3), approximately double the levels observed in the dry season (5.97 ± 2.4%, n = 3). BhC, on the other hand, exhibited minor seasonal differences in organic matter, ranging from 5.6 ± 0.5% (n = 3) to 6.06 ± 1.28% (n = 3) (Fig. 2A, Table S1).

In sediments the redox potential, both in CGSM and BhC, reducing conditions were recorded with a range of values from 28 to 77 mV between both sectors. Increases in redox potential were observed in BhC (52 ± 3, n = 3 to 65 ± 4, n = 3), and decreases in CGSM (50 ± 11, n = 3 to 35 ± 4, n = 3) from rainy season to dry season (Fig. 2A, Table S1).

The silt content showed marked differences between ecosystems and across climatic seasons. In CGSM, average silt content increased notably from 19.95 ± 8.68% (n = 3) during the rainy season to 44.33 ± 11.94% (n = 3) in the dry season, indicating greater fine particle accumulation in the latter period. In contrast, Cispatá Bay exhibited consistently higher silt proportions, with averages of 66.40 ± 5.76% (n = 3) in the rainy season and 66.62 ± 12.14% (n = 3) in the dry season, reflecting stable conditions favorable for fine sediment retention (Table S1).

Concentration of Hg in sediments, seston and C. rhizophorae

Hg concentration in sediments and seston varied markedly between the two ecosystems. In BhC, in both sediments and seston, Hg concentrations are consistently higher than in CGSM in both climatic seasons (Fig. 2B, Table S2).

In the rainy season, the highest concentration of Hg in sediments was found in CIS-1 (BhC) with 0.128 µg/g Hg dry weight (d.w.) which is double the highest content detected in CGSM (0.059 µg/g Hg_d.w. in CGS-1). Hg content in sediments at BhC was slightly lower in the dry season but remained above 0.08 µg/g Hg_d.w. indicating a possible constant source of Hg contamination. In CGSM, during the dry season, lower Hg was observed in CGS-1 and higher in CGS-2 (Fig. 2B, Table S2).

The Hg available in the seston presented similar values in the stations of each ecosystem and in the two climatic seasons. However, as in sediments, a lower concentration was detected in the dry season. In BhC, the concentration went from 0.032 ± 0.005 µg/g Hg_d.w. in the rainy season to 0.022 ± 0.001 µg/g Hg_d.w. in the dry season. Lower concentrations were found in CGSM, with values ranging from 0.01 ± 0.001 µg/g Hg_d.w. in the rainy season to 0.004 ± 0.001 µg/g Hg_d.w. in the dry season (Fig. 2B, Table S2).

In BhC the highest Hg content in seston was positively and significantly related to temperature (Pearson, r = 0.933, df = 11, p-value = 0.006) and in sediments Hg was significantly related to organic matter (Pearson, r = 0.845, df = 11, p-value = 0.039) (Fig. 2, Table S3).

Hg concentrations in oyster tissue show distinct accumulation patterns in CGSM and BhC, varying as a function of climatic seasons. However, a pattern similar to that of sediment and seston was maintained, with a higher Hg content in oyster soft tissue in BhC (Fig. 2B). In the rainy season, Hg in the oyster was 0.083 ± 0.007 µg/g_d.w. (n = 17) in CGSM and of 0.135 ±0.015 µg/g_d.w. (n = 18) in BhC showing significant differences between the two ecosystems (Permanova, p-value < 0.01). In dry season, these differences were maintained, given the decrease in oyster Hg content in CGSM (0.066 ± 0.007 µg/g Hg_d.w., n = 18) and increased in BhC (0.154 ± 0.019 µg/g Hg_d.w., n = 18) (Tables 1 and 2, Fig. 3A, Table S4).

With respect to Hg BCF, both in sediments and seston, both ecosystems presented accumulation to hyperaccumulation of Hg in the oyster tissue, with the highest values in the dry season. In this same climatic season, at CGSM, the oyster presented an accumulation of Hg with the sediment (BCF≥1) and a hyperaccumulation of the metal with the seston (BCF≥10), as opposed to the accumulation condition in both matrices during the rainy season. In BhC, the oyster maintained the accumulation condition in both matrices (BCF≥1) as in rainy season (Tables 1 and 2). These results notably emphasize the capacity of the oyster to accumulate Hg in its tissues, especially in CGSM through the seston in the dry season. Significant differences in BCF were determined between CGSM and BhC ecosystems, with higher concentrations in BhC (Permanova, p-value < 0.05). However, no significant differences in concentrations were observed between climatic seasons, since the values measured at two of three stations in both CGSM (CGS-2 and CGS-3) and BhC (CIS-1 and CIS-3) were similar in both climatic seasons (Tables 1 and 2, Fig. 3B).

In the two ecosystems evaluated, there were significant differences in the Hg contents between stations (Permanova, p-value < 0.05; Table 2). In the rainy season at CGSM the differences occurred between stations CGS-1 and CGS-2 and at BhC between CIS-2 and CIS-3. In the dry season, significant differences were found between CIS-1 and CIS-2 in BhC, with the lowest concentrations in CIS-1 (Table 1, Fig. 2B, Tables S4 and S5).

There were significant differences in the length of juveniles and adults (Permanova, p-value < 0.05) between stations (Table 2). In CGSM in the dry season, concentrations were higher in adult ages at station CGS-3 (0.121 ± 0.016 µg/g Hg_d.w.) with respect to juveniles (0.039 ± 0.009 µg/g Hg_d.w.). In BhC, in both rainy and dry seasons, the highest concentration of juvenile lengths was at station CIS-2 (Table 1, Fig. 3A, Tables S4 and S5).

Importance of seston in the bioconcentration of Hg

High Hg bioconcentration factor values reflected a significant correlation with seston content (Pearson, r = 0.718, df = 11, p-value = 0.008), with conditions of accumulation (BCF≥1) in BhC where an average concentration between both climatic periods was observed in the seston of 0.027 ± 0.003 µg/g Hg_d.w. (n = 6) and in the oyster of 0.144 ± 0.012 µg/g Hg_d.w. (n = 36). While in CGSM hyperaccumulation conditions (BCF≥10) were reached with an average content in the seston of 0.007 ± 0.001 µg/g Hg_d.w. (n = 6) and in the oyster of 0.074 ± 0.005 µg/g Hg_d.w. (n = 35) (Table 1, Table S4).

Differences in Hg bioconcentration between CGSM and BhC were determined (Permanova, p-value < 0.05; Table 2). These differences were also observed as a function of climatic seasons, with an increase during the dry season in each CGSM season, and significant in the CIS-2 season in BhC compared to the rainy season. When the factors ecosystem and climatic season were combined, significant differences were still present, with higher values of Hg bioconcentration in CGSM in both climatic seasons compared to BhC (Table 1, Fig. 3B, Table S5). These results indicate that the oyster in CGSM is accumulating higher concentrations of Hg in its tissues compared to the BhC oyster, although the accumulation is also considerably higher in the BhC oyster.

Significant differences between juvenile and adult ages were determined with the Hg BCF, which was maintained when considering the climatic season (Table 2). In BhC, the highest values of Hg BCF were observed in juvenile ages in both climatic seasons. In CGSM, the highest BCF occurred in adult ages during the dry season, while they were similar in both ages during the rainy season (Spearman, r = 0.25, p-value > 0.05; Table 1, Fig. 3B, Table S4).

Relationship between environmental variables and Hg bioconcentration in oysters

Between the Hg concentration in the mangrove oyster tissue and its BCF by oyster in relation to the metal content in the seston, it was not possible to find a positive or negative relationship with the environmental variables analyzed. The relationship between physicochemical variables with Hg concentration and BCF in the mangrove oyster were not significant (DistLM; p-value > 0.05; Table 3). This suggests that environmental variables did not play a determining role in the differences in oyster Hg content and bioconcentration at CGSM and BhC (Fig. 3). Other factors, such as Hg content in the seston and local transport and sedimentation processes, may be playing a more influential role in Hg accumulation.

Hg contamination status of bivalves in a global context during the last five decadas

Based on a global assessment of total mercury (Hg) contamination in bivalves over the past 50 years across 79 study sites, risk levels were analyzed using the adapted Nemerow Pollution Index (P_c) based on the provisional tolerable weekly intake (PTWI) of one µg Hg/kg body weight. The Ciénaga Grande de Santa Marta (CGSM) ecosystem, evaluated in 2021 and 2022 with the oyster C. rhizophorae, falls within Clade 4. Although this clade includes sites exceeding the very high-risk threshold (P_n >3), it represents the cluster with the lowest relative risk among the four clades. CGSM exhibited a P_c value of 4.77, which, despite surpassing the threshold, remains far below the highest global values. CGSM clustered near sites such as Salerno (Italy, 1987) and Daksa (Croatia, 1984) with the mussel Mytilus galloprovincialis, and Chiayi (Taiwan, 1973) with the oyster Magallana gigas, highlighting a grouping associated with moderate mercury contamination. It is important to note that CGSM exhibited a higher risk of contamination compared to sites such as Saldanha Bay (South Africa, 2016) with Choromytilus meridionalis and M. galloprovincialis, Galicia (Spain, 2006 and 2010) with the scallop Mimachlamys varia, and Changseon Coast (South Korea, 2008) with M. galloprovincialis, all of which remained below the very high-risk threshold with P_n values of 2.33, 2.75, and 2.89, respectively (Fig. 4, Tables S6 and S7).

In contrast, the Cispatá Bay (BhC) ecosystem, assessed in 2021 and 2022 with C. rhizophorae, recorded a P_c value of 5.92 and was grouped within Clade 3, positioning it in the third-highest contamination cluster. BhC clustered with sites such as the Yellow Sea and Bohai Sea (China, 2019) with Mactra veneriformis, Chlamys farreri, Ruditapes philippinarum, M. gigas, and Cyclina sinensis; the Sibenik Aquarium (Croatia, 1983) and Piombino (Italy, 1987) also with M. galloprovincialis. This indicates a higher relative risk compared to CGSM, but still distant from the extreme contamination observed in Clade 1, which includes heavily impacted sites such as Kastela Bay (Croatia, 1980–1985) with M. galloprovincialis, the “Cheminova” factory area (Denmark, 1982–1983) with M. galloprovincialis and Macoma balthica, and the English Channel (United Kingdom, 1975) with Mytilus edulis. In these locations, Hg concentrations in bivalves exceeded the PTWI limit of one µg Hg/kg body weight established by the World Health Organization (WHO) by more than an order of magnitude (Fig. 4, Tables S6 and S7).

These results highlight that although both Colombian sites exceed the PTWI threshold and thus warrant attention, they are far from the global hotspots of mercury contamination in bivalves. Instead, they are grouped within clusters representing intermediate pollution levels when analyzed in a historical global context.

Hg intake risk in C. rhizophorae at CGSM and BhC compared to global hotspots

This study offers a critical contribution to the understanding of total Hg pollution in C. rhizophorae within the Colombian Caribbean, specifically in the ecosystems of Ciénaga Grande de Santa Marta (CGSM) and Cispatá Bay (BhC). Among 79 study sites spanning 36 bivalve species and 32 global investigations from 1970 to 2022 (Fig. 4, Table S6), CGSM and BhC are the only locations where C. rhizophorae has been assessed for Hg contamination using wet weight concentrations (µg/kg). These data are standardized against the provisional tolerable weekly intake (PTWI) of one µg Hg/kg body weight established by FAO/WHO (2010), allowing for robust and consistent comparison of Hg exposure risk through seafood consumption.

Although Hg concentrations in C. rhizophorae from CGSM (16.7 ± 1.9 µg/kg_w.w., n = 35) and BhC (29.2 ± 3.2 µg/kg_w.w., n = 36) exceeded the PTWI threshold, these levels remain significantly lower than those found in ecosystems classified in Clade 1, which includes historical global hotspots of extreme contamination. For instance, M. galloprovincialis from Kastela Bay (Croatia) reached 9,525 ± 2,337 µg/kg_w.w. (Mikac et al., 1985), while M. balthica near the “Cheminova” factory (Denmark) recorded 1,260.3 ± 119.6 µg/kg_w.w. (Riisgard et al., 1985). Other examples include S. plana from the Ria de Aveiro (Portugal) with 395 ± 87 µg/kg_w.w. (Coelho et al., 2014), and Tegillarca granosa from East Java, Indonesia with 272.5 ± 172.1 µg/kg_w.w. (Soegianto et al., 2020) (Fig. 4, Tables S6 and S7).

While these international assessments have traditionally emphasized species like M. edulis and M. galloprovincialis monitored in more than 10 countries such as Italy (Brambilla et al., 2013; Esposito et al., 2021), Spain (Olmedo et al., 2013), and South Africa (Firth et al., 2019), the representation of C. rhizophorae in global or regional Hg monitoring is notably scarce. Our study fills this critical gap by generating the first standardized dataset for this ecologically and commercially important Caribbean species, evaluated through the lens of direct human exposure risk. In fact, most historical studies on C. rhizophorae have used dry weight units, as seen Dominican Republic (Sbriz et al., 1998), Brazil (Silva, Rainbow & Smith, 2003), and the Nicaragua (Aguirre-Rubí et al., 2017), which complicates direct risk assessment for consumption compared to the approach used here.

This highlights not only the regional novelty of the current study but also the broader need for consistent wet weight-based assessments of Hg in seafood species of local significance. Moreover, by implementing an adapted Nemerow Pollution Index standardized for PTWI, this study advances a globally comparable metric of consumer health risk one that remains absent from many other international assessments.

Nevertheless, the current state of Hg contamination in Colombia, particularly in Cartagena Bay, highlights the need for vigilance. Cartagena Bay has recorded some of the highest global Hg contamination risks for bivalves in the past decade (Aguirre-Rubí et al., 2017), attributed to historical discharges from industrial facilities like the Alcalis chlorine plant (Mancera-Rodríguez & Álvarez-León, 2006; Bolaños-Alvarez et al., 2024; Ucros-Rodríguez et al., 2025). Similar trends have been observed in other regions, including “Cheminova” factory, Denmark (Riisgard et al., 1985), Kastela Bay, Croatia (Mikac et al., 1985; Martinčić et al., 1987) San Vicente Bay, Chile (Díaz et al., 2001), California Gulf, United Stades (Ruelas-Inzunza, Soto & Páez-Osuna, 2003), and the Adriatic Sea, Croatia (Kljaković-Gašpić et al., 2010), reinforcing the global relevance of monitoring mercury in estuarine and coastal ecosystems. In this context, the current data from CGSM and BhC establish vital benchmarks that can inform targeted mitigation strategies, ensuring ecosystem protection and seafood safety in the region.