All sediment and seawater samples were collected in November 2021 from the eastern region of the South China Sea (Figure 1). A total of 10 sampling points were selected, namely, W1/S1, W2, W3/S3, W4, W5, W6/S6, W7, W8, W9, and W10/S10. Here, “W” denotes water, and “S” denotes sediment. The sampling point W1/S1 was located in a densely populated area that had undergone a dredging project. Points W3/S3 and W4 were situated in close proximity to an aquaculture zone, while W6/S6 was situated near Guangdong province’s renowned aquaculture area. Finally, points W9 and W10 were situated far from shore. The sediment samples were collected using an SDI VC-D sediment sampling rig. The samples were then immediately cut into 2 cm pieces, sealed in zipper storage bags, and stored in a refrigerator. Seawater samples (from about 50 cm below the water surface) were collected using a stainless steel container and stored in 2-L polypropylene (PP) containers. In total, four surface sediment samples, two sediment cores, and 10 seawater samples were collected and transported to the laboratory before being stored at −20°C. All tools and containers were cleaned at least three times with Milli-Q water and methanol. The information on the sampling sites is listed in Supplementary Table S1.

A total of 21 PFAS standards (PFBS, PFHxS, PFOS, PFDS, PFBA, PFPeA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA, PFDoA, PFTrDA, PFTeDA, PFHxDA, PFODA, PFMOPrA, PFMOBA, ANODA, and F-53B) and 9 PFAS internal standards (¹³C₄PFBA, ¹³C₄ PFHxA, ¹³C₄ PFOA, ¹³C₄ PFNA,¹³C₄ PFDA, ¹³C₄ PFUndA, ¹³C₂ PFDoA, ¹⁸O₂ PFHxS, and ¹³C₄ PFOS) were purchased from Wellington Laboratories (Guelph, ON, Canada). The internal standards each had purity equal to or exceeding 98%. Acetonitrile, methanol, and methyl tert-butyl ether (MTBE) were purchased from J.T. Baker (Phillipsburg, NJ, United States). Ammonium acetate was purchased from J&K Chemicals. Detailed information is listed in Supplementary Table S2.

Seawater samples were prepared and purified according to the following methods (Wang et al., 2019; Diao et al., 2022). Fiberglass filters (GFF, Whatman, O.D. 47 mm, 0.7 μm) were used to filter the seawater samples. Then, 5 ng of the internal standard was added to an 800-mL sample of seawater before extraction by solid-phase extraction (SPE). Specifically, the OASIS WAX cartridges were treated with 4 mL of 0.1% NH₄OH in methanol, 4 mL of methanol, and 4 mL of Milli-Q water, followed by 25 mM ammonium acetate that was used to wash the cartridge after extraction. The cartridges were eluted using 4 mL of 25 mmol ammonium acetate, 4 mL of methanol, and 4 mL of 0.1% NHOH in methanol. The eluate was concentrated to 0.5 mL using high-purity nitrogen. Finally, a nylon filter was used to filter the concentrated solution into a 1.5-mL PP snap-top autosampler via a polyethylene (PE) septa. The full details are given in the Supplementary Material.

The sediment samples were wrapped in aluminum foil and frozen at −20°C in a refrigerator. After freeze-drying the samples for 48 h, the sediment samples were ground into powder in a grinding machine. Next, about 2 g of the sediment sample, 2 mL of the sodium hydroxide: acetonitrile solution (8:2), and 5 ng of the internal standard were added to a 50-mL PP tube, followed by ultrasonication with an ultrasonic cleaner and the subsequent addition of 20 mL of acetonitrile. After shaking for at least 30 min at 250 rpm, a 100-μL solution of 2 mol/L hydrochloric acid was added. The mixture was then centrifuged at 3,000 rpm for 15 min, and the organic layer was separated into a new PP tube. This solution was then evaporated using a nitrogen stream. The ENVI-Carb was cleaned three times with 1 mL of methanol, and the concentrated solution was transferred to the ENVI-Carb. The PP tube was also cleaned with 3 mL of methanol, and the solution was transferred to the ENVI-Carb. The following steps were the same as those of the SPE method (Chen et al., 2019; Sun et al., 2021a).

The biological samples were dissected and freeze-dried at −50°C for 72 h. A grinding machine was used to grind the muscle tissue into powder. Approximately 1 g of the biological sample was added into a 15-mL PP tube along with 1 mL of 0.5 M TBAHS, 2 mL of Na₂CO₃ aqueous solution, 5 ng of the internal standard, and 5 mL of MTBE. The remaining steps are the same as those used to analyze the sediment samples (Sun et al., 2021b).

A Thermo Ultimate 3,000 Infinity HPLC System equipped with a Thermo TSQ ENDURA LC/MS System was used for electrospray ionization (ESI) in the negative ion mode with multiple reaction monitoring (MRM) of the target analytes. The HPLC-ESI-MS instrument conditions are listed in Supplementary Table S3. An Agilent ZORBAX Eclipse Plus (C18, 2.1 × 100 mm, 3.5 μm, Agilent) connected with a guard column (Agilent) was used as the column. The mobile phase used was 2 mM of ammonium acetate (A) and 100% acetonitrile (B). Then, 5-μL samples were extracted and injected into the column at 0.3 mL/min. Supplementary Table S4 lists the chemical formulas, parent ions, quantitative ions, and qualitative ions of the PFAS tested in this study.

The sediment core samples were analyzed for radionuclides at the State Key Laboratory of Estuarine and Coastal Sciences, East China Normal University (Shanghai, China). The ²¹⁰Pb and ¹³⁷Cs radioactive dating techniques were used to determine the activities of ²¹⁰Pb, ¹³⁷Cs, ²²⁶Ra, and other radioactive isotopes in each of the sediment samples (Appleby, 2008; Zhang Y. Q. et al., 2021). The specific operation steps were as follows: a freeze-dried and ground sample was placed in a clean PP tube for compaction and testing after 1 month. The samples were analyzed using gamma spectrometry by counting for at least 24 h. The excess ²¹⁰Pb (²¹⁰Pbex) activity in the sample was determined from the difference between the total ²¹⁰Pb activity and ²²⁶Ra activity.

After drying the soil sample, approximately 5 g were retrieved to remove impurities (Lund et al., 2021). Chemical pretreatment was then performed before testing: (1) the samples were each treated with 30% H₂O₂ for 24 h to remove organic matter; (2) to remove carbonate impurities, 10% HCl was used for 24 h; (3) the particle size distribution of each sample was measured using the Mastersizer 3,000 (Malvern Panalytical, UK). The average value of each sample was obtained from three repeated measurements, and the guaranteed error was less than 0.1%. The theoretical basis of the Mastersizer 3,000 is the Michaelis scattering theory, with a measuring range of 0.02–2000 μm and granularity resolution of up to 0.01 φ.

To reduce the influence of background contamination, field and laboratory blanks were prepared. Twenty samples were used to prepare a blank, which was then analyzed in the same manner as the samples. The results showed that the concentrations in all the blanks were significantly lower than those of the samples. All experimental tools were cleaned at least three times using Milli-Q water and methanol. The limit of detection (LOD) was given by a 3:1 signal-to-noise ratio, and the limit of quantification (LOQ) was determined by a 10:1 signal-to-noise ratio. The rates of recovery for water, sediment, and biota ranged between 80% and 120%. The gradients of the curves were set at concentrations of 0.005, 0.05, 0.1, 1, 10, 50, 100, and 200 ng/mL with 10 ng/mL of the internal standards. Each of the PFAS had a correlation coefficient (R ²) of around 0.99. The LOD, LOQ, and recovery rates are shown in Supplementary Table S5.

In this study, a total of four perfluoroalkyl sulfonic acids (PFSAs: PFBS, PFHxS, PFOS, and PFDS) and 12 perfluorinated carboxylic acids (PFCAs: PFBA, PFPeA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA, PFDoA, PFTrDA, PFTeDA, and PFHxDA) were detected in the water, sediment, and biota samples. Figure 2 displays the molecular structures of the principal PFAS compounds examined. As the concentrations of PFDS were lower than those of MDL, they are not discussed henceforth. Additionally, the investigation revealed the presence of four emerging PFAS compounds (ADONA, PFMOBA, PFMOPrA, and F-53B) in both sediment and water samples. Detailed information on the detection rates (DRs) is listed in Supplementary Table S6.

The concentrations of the PFAS in the water samples ranged from 13.08 to 57.70 ng/L, with an average of 29.60 ng/L (Figure 3A). Notably, higher PFAS concentrations were recorded in samples from the streamway (W3: 57.70 ng/L, W4: 43.83 ng/L) and the bay (W6: 42.06 ng/L) compared to the other locations. Offshore sampling sites (W8: 13.08 ng/L, W9: 20.33 ng/L, W10: 17.18 ng/L) had low concentrations compared to the other sites (Table 1). In the surface sediment samples, the concentrations ranged from 0.96 to 1.35 ng/g dw, with an average of 1.17 ng/g dw. The inshore (S1: 1.35 ng/g dw) and streamway (S3: 1.34 ng/g dw) locations registered the highest PFAS concentrations, whereas the lowest concentration was detected offshore (S10: 1.04 ng/g dw). Interestingly, in the surface sediments of the stream (S3), bay (S6), and offshore (S10) sites, the highest concentrations were not found in the top layer (0–4 cm), but rather in the second layer (4–8 cm), as illustrated in Figure 3B. As the ocean sediment is the ultimate sink for the PFAS, these results provide new evidence for the depletion of PFAS stocks in the surrounding market in the context of gradual phase-out and restriction. The detailed water and sediment concentrations are listed in Supplementary Table S7, and the surface sediment core concentrations are listed in Supplementary Table S8.

S1 and W1 are situated on the east coast of Shantou city, an area affected by both sea reclamation and intense human activity. W3 and W4 are located downstream of the Rong River, which flows through a densely populated area and has the highest concentration of pollutants. In contrast, W2, which is situated near the estuary, benefits from the natural dilution and dispersion processes inherent to such geographic settings, which help mitigate pollutant concentrations. Despite this natural advantage, pollutant levels at W2 remain elevated compared to those in adjacent waters, underscoring the persistent impact of human activity on the environment. W6 is close to the coast and is an offshore aquaculture area, where the abandonment of aquaculture equipment may lead to the continuous release of PFAS. In addition, W6 is located in a bay, which reduces the water exchange capacity of this region. W8, W9, and W10 are situated far from the shore and are subject to dilution by seawater; therefore, the concentrations of PFAS were low. The water and sediment samples showed similar trends, revealing high levels of PFAS in the nearshore areas, which decrease toward open marine waters.

The concentrations in the biota samples ranged from 9.71 to 25.44 ng/g dw, with an average of 15.80 ng/g dw. Notably, higher PFAS concentrations were detected in organisms occupying higher trophic levels, such as Clarias macrocephalus and Portunus gladiator Fabricius, with concentrations of 17.79 ng/g dw and 25.44 ng/g dw, respectively. Mytilus edulis (B2: 10.26 ng/g dw) and Ostrea gigas Thunberg (B3: 9.71 ng/g dw) are in the low trophic level, so their PFAS concentrations were lower than those in B1 and B4 (Supplementary Table S9). When comparing these results to those of other studies, the PFAS levels in the biota from the eastern region of the South China Sea are considered to be moderate to low (Table 1), suggesting a varied distribution of PFAS contamination across different marine environments and species.

In this study, a total of 16 target PFAS compounds, including three novel PFAS (F-53B, PFMOBA, and PFMOPrA), were identified in water samples, as illustrated in Figure 4. The frequencies of detection and concentrations of the long-chain PFAS (C > 8) were lower than those of the short-chain PFAS (C < 8). This trend is attributed to the higher water solubility and lower n-octanol/water partition coefficients of the short-chain PFAS, making them more prevalent in aquatic environments. The higher proportion of short-chain PFAS in the South China Sea was consistent with the observations of previous studies (Zheng et al., 2017; Wang et al., 2019). The dominant compounds were PFBA (49.43%–90.76%; 10.05 ng/L to 46.12 ng/L) and PFOA (3.12%–10.31%; 1.35 ng/L to 2.33 ng/L). Compared with other studies, Wang et al. (2018) reported the presence of PFBA (17.8 pg/L to 340.8 pg/L) and PFOA (< LOQ to 400.7 pg/L), which were the main compounds in the coastal region of the South China Sea. The noted increase in PFBA levels and decrease in PFOA concentrations suggest transitions in PFAS usage from long-chain to short-chain compounds in coastal areas. Specifically, PFPeA was detected exclusively offshore (W9), and PFUnDA was found only in the river samples (W4). Notably, the novel PFAS PFMOBA and PFMOPrA were detected in all water samples, replacing PFOA and PFOS. F-53B was detected at five sites. The toxicological implications of these novel PFAS are not fully understood, and the study by Woodlief et al. (2021) suggests potential effects on the immune system. With the increasing use of novel PFAS, their release into the environment is expected to increase over time. Unlike long-chain PFAS, which can be treated through traditional processes, novel PFAS pose challenges for treatment and have toxicological potential (Zhang et al., 2021; Woodlief et al., 2021).

Seventeen PFAS were detected in the sediment samples: 10 PFCAs (PFBA, PFPeA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA, PFDoA, and PFTrDA), 3 PFSAs (PFBS, PFHxS, and PFOS), and 4 novel PFAS (F-53B, ADONA, PFMOBA, and PFMOPrA). In contrast to the water samples, PFTrDA and ADONA were detected and PFHxDA was not found in the sediment samples. Similar to the findings in the water samples, short-chain PFAS were more frequently detected in the sediments than long-chain PFAS, indicating a significant transfer of short-chain compounds from water to sediment. Notably, PFTrDA was detected in the surface sediment at a streamway sampling site (S3); the deposition process at this site is known to be influenced by external factors and is also possibly impacted by shipping and human activities. Sediments are recognized as the ultimate sink for PFAS compounds with eight or more carbon atoms owing to their higher n-octanol/water partition coefficients (Prevedouros et al., 2006). Despite the prohibition of long-chain PFAS, which limits their presence in water samples, they were still identifiable in the sediment samples. PFBA (73.97%; 0.78–0.99 ng/g), PFHxA (3.58%; nd. to 0.15 ng/g), and PFOS (3.05%; nd. to 0.12 ng/g) were the dominant compounds. Similar results were reported by Diao et al. (2022), who found that short-chain PFAS had higher DRs than long-chain PFAS in sediments. However, Mourier et al. (2019) reported contrasting results, which may be caused by the discharge of short-chain PFAS along the coast of the South China Sea. A significant amount of the short-chain PFAS was then rapidly deposited from the water into the sediment. In the surface sediment, PFBA was still the major compound, while PFDA, PFUnDA, and PFOA were the dominant long-chain PFAS (Supplementary Figure S1). Notably, the layers extending from 4 to 12 cm depth in the streamway (S3), bay (S6), and offshore (S10) sites exhibited higher concentrations of the long-chain PFAS, indicating their extensive use in the past. Interestingly, high concentrations of long-chain PFAS were observed at the inshore site (S1) in the 0–4 cm and 16–20 cm layers, possibly due to sediment layer disturbances from past reclamation projects. Similar to the water samples, PFMOPrA, a novel PFAS, was found in the surface sediments from all sites, indicating that PFMOPrA was rapidly transported from the water to the sediment.

Sixteen PFAS were detected in the biota samples: 12 PFCAs (PFBA, PFPeA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA, PFDoA, PFTrDA, PFTeDA, and PFHxDA), 3 PFSAs (PFBS, PFHxS, and PFOS), and 1 novel PFAS (F-53B). Among these, PFOS, PFBA, and PFUnDA were the predominant compounds in the biota samples, with concentrations ranging from 0.32 to 12.42 ng/g dw for PFOS, 2.14 to 5.20 ng/g dw for PFBA, and 0.07 to 5.19 ng/g dw for PFUnDA. As seen in Supplementary Figure S2, the concentrations of the short-chain PFAS in M. edulis (B2) and O. gigas Thunberg (both filter feeders) were higher than those in the other two biota samples. In contrast, C. macrocephalus (B1, long-chain PFAS concentration: 12.53 ng/g dw) and P. gladiator Fabricius (B4, 22.92 ng/g dw) as the omnivores had higher concentrations of the long-chain PFAS. It is likely that filter feeders source their food from water-based plankton, while omnivores source their food from marine plants and animals. On the other hand, marine flora and fauna may contribute to the accumulation of long-chain PFAS in omnivores, explaining their higher levels of long-chain PFAS. Filter feeding may thus result in elevated levels of short-chain PFAS in the seawater.

The examination of sediment core samples from inshore (S1) and offshore (S10) locations in this study provides an understanding of the spatial distribution and temporal accumulation of PFAS in marine sediments. The length of the S1 sediment core is 120 cm, and the length of the S10 sediment core is 146 cm. The concentrations of PFAS in the S1 sediment core ranged from 0.54 to 2.34 ng/g, and those in the S10 sediment core were in the range of 0.43–1.17 ng/g. The median grain sizes ranged from 5.86 to 93.63 μm in S1 and from 5.80 to 7.56 μm in S10. In S1, there were no significant changes in grain size at a depth of 0–70 cm, but the grain sizes beyond 70 cm were significantly larger than those above, with the maximum grain size observed at 90–94 cm. In S10, the grain size showed a gradually decreasing trend, with a minimum value of 120–122 cm. Overall, the values of the PFAS concentrations and grain sizes in the sediment core of S1 were higher than those of S10. Larger grain sizes may provide more surface area for the adsorption of PFAS, leading to higher concentrations in certain sediment layers. The concentrations of PFAS and grain sizes in the sediment cores are listed in Supplementary Table S10.

The deposition dates of the sediments were determined based on vertical changes to ¹³⁷Cs and sediment grain sizes. For the S1 sediment core, the observation that the surface layer (0–12 cm) was significantly eroded by external forces complicated the dating process, indicating that this layer was rapidly deposited over the last 10 years but cannot be dated with precision. The rest of the S1 sediment core, however, could be accurately dated, revealing a deposition rate of 0.94 cm/year. This rate allows chronological reconstruction of the sediment deposition back to the year 1910, providing over a century’s worth of environmental pollution data. The chronological determination of the S10 sediment core accurately reflects the age of each layer, with a deposition rate of 1.24 cm/year. Accurate dating of the S10 layers extends even further back in time, to the year 1880, thus offering a longer historical perspective on sediment deposition and the associated environmental changes.

As shown in Figure 5, the vertical profile of the sediment and its division into three distinct segments based on the analysis of PFAS concentrations provide a compelling narrative of the environmental impacts of industrial development, policy changes, and environmental management over time. This analysis offers insights into the temporal dynamics of PFAS pollution as well as its relationship with socioeconomic developments and regulatory interventions.

(1) First segment: Policy limitation period (2000s–2020s). During this period, many countries successively implemented policies to restrict the use of PFAS, including the restrictions placed by the Chinese government on long-chain PFAS, such as PFOS and its related compounds, as listed in the Stockholm Convention. Consequently, the concentration of PFAS in S10 showed a gradual increase in the 2000s, followed by a sharp decline after reaching a peak in the early 2010s. In S1, the peak occurred in the early 2000s and then decreased gradually. After 2010, the concentration of PFAS in S1 remained low; this can be attributed to the influence of local human activities and land-based pollution sources, particularly in areas with high industrial activity or significant land development projects. Compared with S1, the concentration peak of PFAS in S10, which is far from land, has a hysteresis, which means that the PFAS concentration in the sediment is predominantly influenced by land-based pollution. From this, it can be seen that the concentrations of PFAS have been decreasing steadily during the policy period. It is expected that the concentrations of PFAS in the environment will decrease further and that the impacts of PFAS on the environment and humans will be mitigated after the implementation of the new pollution control action plan by the Chinese government in 2023.

The use of PCA and cluster analysis in this study provides a nuanced understanding of the origins, distributions, and behaviors of PFAS compounds across different environmental matrices (sediment, water, and biota) and geographical locations (inshore and offshore). These analyses help dissect the complex environmental dynamics of PFAS pollution, highlighting the interplay between historical industrial discharges, current emissions, and the natural environmental processes that redistribute these pollutants.

PCA was used in this study to explore the origins of different PFAS compounds (Xiao et al., 2021) and to determine the similarities among the sampling sites (Figure 6A). PC1 and PC2 accounted for 39.81% and 21.34% of the total variations, respectively. PC1 was identified as a long-chain PFAS due to its close relationship with PFNA, PFOS, PFDA, PFUnDA, PFDoA, PFTeDA, PFTrDA, and PFHxDA, with the exception of F-53B, which is classified as a novel PFAS. Long-chain PFAS have been banned in China, and all of the identified long-chain PFAS in PC1 may have therefore come from the residues of past industrial discharges. Owing to the phase-out of long-chain PFAS, short-chain PFAS have been widely used as substitutes. Notably, PFOA removal rates in wastewater treatment plants were negative, and it was observed that fluorotelomers can transform into PFOA via the treatment process. The PFAS components identified in PC2 may originate from domestic sewage and industrial wastewater. This is similar to the previous findings that PFOA, PFOS, PFHxS, PFBA, and PFHxA were the most dominant PFAS compounds found in domestic effluents (Zhang et al., 2021; Zhou et al., 2022). PFPeA and PFHxS were significantly negatively correlated with PC2. Zhang et al. (2016) reported that PFPeA was the major component in the Beijing atmosphere. It is inferred that PFPeA and PFHxS may have originated from atmospheric sedimentation. Moreover, water samples are clustered in the left part of the PCA, and sediment and biota samples (B2: M. edulis and B3: O. gigas Thunberg) are clustered in the lower part of the PCA. These results suggest that there were differences in the PFAS concentrations between the water, sediment, and biota samples. However, the results of the sediment and biota samples (B2 and B3) were found to be relevant; this may be due to the filter-feeding mode, which enriches the PFAS from the sediments, and the benthic life affected by the sediments.

Cluster analysis was used in this study to categorize the PFAS and determine the origins of their various components (Chen et al., 2021; Mu et al., 2022). As shown in Figure 6B, the PFAS were clustered into three groups. C1: The isolation of PFBA in its own category suggests unique sources or behaviors distinct from other PFAS compounds. This could imply specific industrial or consumer product-related emissions. C2: Containing PFOS, PFHxA, PFHpA, PFNA, PFOA, and PFBS, it likely represents the PFAS with common industrial applications. C3: Containing PFHxS, PFPeA, F-53B, PFDA, PFUnDA, PFHxDA, PFDoA, PFTeDA, and PFTrDA. This classification was similar to that of PCA. Additionally, the sample points were divided into three distinct groups. The first group (G1) consisted of biota and sediment samples. The second group (G2) comprised water samples (W1, W2, W5, W7, W8, W9, and W10), which were further divided into two subgroups (inshore: W1, W2, W5, and W7; offshore: W8, W9, and W10). Water samples W3, W4, and W6 were assigned to the third group (G3).

The results of this study show that areas close to the coast (inshore) are significantly affected by pollution due to activities of nearby human populations and industries. Nan’ao Island, which is located near these coastal areas, acts as a natural barrier, and this barrier effect can prevent pollutants from dispersing effectively into the open ocean. As a result, pollutants, including harmful chemicals, tend to accumulate in these inshore areas, leading to higher levels of pollution compared to areas farther from the coast (offshore). In contrast, the sampling sites located offshore and away from land did not experience any barrier effects from the island; hence, the offshore group displayed significantly lower pollutant concentrations than the other sites. This was consistent with the horizontal transport pattern in the ocean, which spread from nearshore regions with high human activity to more remote seas.

The RQ is a common index used to assess ecological risk. The MEC and PNEC are the measured environmental concentration and predicted no-effect concentration, respectively. A value of RQ >1 indicates high potential risk, 0.1 < RQ ≤ 1 represents moderate potential risk, and RQ <0.1 suggests minimal potential risk. The PNEC values are listed in Supplementary Table S11 (Pan et al., 2014; Xiao et al., 2021). PFOA and PFOS have the potential to pose toxic effects on various trophic levels of organisms, while seawater and sediments have the capacity to directly or indirectly affect marine biota. Therefore, it is necessary to understand the environmental risks of sediment and seawater PFOA and PFOS aquatic systems. According to Figure 7A, the RQ values of sediment samples range from 0.032 to 0.259 for PFOS and from 2.9 × 10⁻⁴ to 2.3 × 10⁻³ for PFOA. The results of the sediment samples show that most of the surveyed areas had negligible risks, with RQ values <0.1. Only one inshore sediment sample (S1: RQ = 0.259) showed a medium risk. For the water samples, the values of RQ were <0.1 for all PFAS compositions. In the water and sediment sampling areas, the RQ of a mixture of 10 PFAS was less than 0.1, except for S1, indicating that the PFAS in marine waters and sediments east of Guangdong pose minimal potential risks. Currently, PFAS do not cause serious risks to the aquatic ecosystem of the coastal area of eastern Guangdong, but their long-term risks (high bioaccumulation and biomagnification) cannot be ignored (Du et al., 2021; Diao et al., 2022).

The absence of significant ecological risks from PFAS in the ocean does not imply the absence of enrichment within marine biota. Figure 7B shows the BAF, a crucial parameter for analyzing the partitioning behavior of organic contaminants in aquatic ecosystems (Du et al., 2021). The BAFs of eight PFAS (PFBA, PFHxA, PFNA, PFDA, PFOA, PFOS, PFHpA, and F-53B) were calculated in this study. Both P. gladiator Fabricius and C. macrocephalus are top predators that demonstrate significant bioaccumulation of long-chain PFAS. This observation implies that PFAS concentrations in these apex predators can increase substantially through trophic transfer, as supported by the findings of Diao et al. (2022) and Xiao et al. (2021). The novel PFAS F-53B also shows bioaccumulation in P. gladiator Fabricius, suggesting that humans may be ingesting new PFAS through the enrichment of marine biota. These findings suggest that top predators tend to accumulate long-chain PFAS. Therefore, it is advisable to limit the consumption of such seafood species. While direct PFAS accumulation in the marine biota was not deemed significant, the potential for bioaccumulation and its implications on both marine organisms and human health warrants careful consideration. BAF values are listed in Supplementary Table S12.

A small number of top predators showed enrichment effects for long-chain PFAS, which could lead to PFAS accumulation in humans through consumption of these seafood species and thereby pose risks to human health. To quantify this risk, the EDI and HR methods are utilized in this work, focusing primarily on PFBA, PFOA, PFHxA, PFBS, PFOS, PFHxS, PFNA, and PFDA. These metrics offer insights into the potential health risks posed by PFAS exposure across different age and gender demographics. A value of HR >1 indicates that the current PFAS exposure may pose a potential risk to humans. As shown in Supplementary Table S14, the EDI values for the eight PFAS ranged from 0.13 to 7.65 ng/kg⋅bw/day. The EDI order among PFAS is as follows: PFOS > PFBA > PFDA > PFOA > PFHxS > PFNA > PFBS > PFHxA. Notably, adolescents (aged 15–18 years) exhibited the highest EDI values, likely due to increased food consumption during their rapid growth phase. In the maturity group, the value of EDI was higher for men than women. The values of HR ranged from 3.07 × 10⁻⁴ (PFBS) to 3.82 (PFOS). Except for PFOS, the HR values of the other seven PFAS were less than 1. This was consistent with the findings of a previous study in the Beibu Gulf and Yellow Sea (Zhang et al., 2016). The order of HR was as follows: PFOS > PFOA > PFNA > PFDA > PFHxS > PFBA > PFHxA > PFBS. The particularly vulnerable groups to PFAS exposure were those in the age ranges of 6–10 and 15–18 years, with men facing a higher risk than women. Supplementary Tables S13–S15 outline these EDI and HR data.