Consistencies and differences between MSFD requirements and ESB standards
The suitability of the ESB samples for D9 assessment was evaluated with respect to the MSFD requirements. Table 1 summarizes the main results. More details are given in Tables S1 and S2 (Supporting Information).
Geographical and temporal scope
D9 monitoring requires that “the temporal and geographical scope of sampling is adequate to provide a representative sample of the specified contaminants in fish and other seafood in the marine region or subregion” (EC 2017a). The ESB samples cover three coastal regions that are considered representative of German coasts. However, the fishing grounds in the open seas of the EEZ are not covered. With respect to the temporal scope, the ESB sampling of blue mussels meet the D9 requirements. In the case of eelpout, the samples are not representative for the whole year but may be considered as worst-case scenario with respect to lipophilic contaminants in fillets because sampling takes place prior to mating season and thus before body lipids are transferred to the reproductive tissue (Greenfield et al. 2005).
The MSFD does not predefine the fish or mussel species and sizes to be analyzed for D9 assessment. The only specifications are that the species has to (1) be relevant to the marine region, (2) fall under the scope of Reg. (EC) No. 1881/2006, (3) be suitable for the contaminants, and (4) be among the most consumed in the Member States or the most caught or harvested for consumption. An indicative list with species is included in Swartenbroux et al. (2010). Zampoukas et al. (2014) also recommends to consider the ability of the species to biomagnify specific classes of contaminants and to ensure that different trophic levels and habitats are represented.
Blue mussels meet all the above mentioned criteria: They are frequent inhabitants of the German coastal regions and are commercially exploited for human consumption. The mussels live attached to rocks, poles, and other hard substrates in intertidal areas. As filter feeders they are exposed to both water soluble and particle bound contaminants and are therefore widely used as bioindicators and chemical pollution monitoring species in coastal waters (review: Beyer et al. 2017).
The suitability of eelpout for D9 assessment is less obvious. Eelpouts are abundant demersal fish in the German coastal regions. They have low migratory behavior and are often used as coastal bioindicators for biological and chemical effect assessments (Hedman et al. 2011; HELCOM 2017b; OSPAR 2013). However, while they are a welcome by-catch in commercial fisheries, they are not specifically fished upon and are not among the most caught or consumed species in Germany (listed in Centenera 2014).
Nevertheless, the contamination of eelpout can give an indication of the contamination of some of the most consumed fish because of similarity of exposure and habitat. Eelpout feed mainly on bottom-dwelling organisms like snails, insect larvae, crustaceans, and eggs and fry of fish. Their trophic level (TL) is around 3.5 which makes them comparable to, e.g., plaice (Pleuronectes platessa, TL 3.2) and flounder (Platichthys flesus, TL 3.3; all TL data according to Froese and Pauly 2017). Like plaice and flounder, eelpouts live in close contact to the sediment and are exposed not only to contaminants in the water phase and to bioaccumulated contaminants via trophic transfer but also to substances bound to the sediment.
The suitability of eelpout for D9 assessment is supported by data from Karl and Lahrssen-Wiederholt (2009) and Karl et al. (2010) who analyzed PCDD/Fs, dl-PCBs, and ndl-PCBs in cod (Gadus morhua) and herring (Clupea harengus) from georeferenced sites in the North and Baltic Seas: After lipid normalization levels in eelpout, cod and herring differ by no more than a factor of 2.
Sampling and sample processing
According to Commission Decision (EU) 2017/848, sampling for the assessment of the maximum levels of contaminants (D9) shall be performed in accordance with the quality standards required in food legislation laid down in Reg. (EC) No. 882/2004 (regarding the performance of controls to ensure compliance with feed and food law, animal health and animal welfare; EC 2004), Reg. (EU) No. 644/2017 (regarding sampling and analysis for the control of dioxins, dl-PCBs and ndl-PCBs; EC 2017b), and Reg. (EC) No. 333/2007 (regarding sampling and analysis of metals and B[a]P; EC 2007 amended by EC 2011d).
The ESB meets most requirements regarding sampling and analysis for the control of dioxins, dl-PCBs, ndl-PCBs, and metals or has even stricter standards. Details are given in Tables S1 and S2 (Supporting material).
However, sampling under the ESB is not performed by officially authorized personnel according to Reg. (EC) No. 333/2007 and Reg. (EC) No. 644/2017, and no controls according to Reg. (EU) No. 882/2004 are performed during sampling and processing.
In the case of blue mussels, it should also be kept in mind that ESB mussel samples include the breathing water which accounts for in average 58% of the mussel wet weight at NS 2 and 67% at NS 1 and BS. Accordingly, respective factors of 3 (NS 1), 2.4 (NS 2), and 3 (BS) should be considered when comparing mussel data with maximum levels (MLs).
Analyzed contaminants and chemical analyses
D9 refers to contaminants in fish and other seafood for which regulatory levels have been set to protect human consumers. These are heavy metals (Pb, Cd, Hg), PAHs (benzo[a]pyrene, benzo[a]anthracene, benzo[b]fluoranthene, and chrysene), dioxins, furans and dioxin-like PCBs (PCDD/Fs + dl-PCBs), and non-dioxin-like PCBs (ndl-PCBs).
Currently, all D9-relevant contaminants are covered by the ESB. The substances are analyzed either in blue mussel or eelpout samples (or in both). For PAHs, however, the analytical method used within the ESB program might lead to an overestimation regarding benzo[b]fluoranthene and chrysene to which the regulatory levels apply (EC 2011c) because the ESB’s analytical method is not able to distinguish between benzo[b]fluoranthene, benzo[j]fluoranthene, or benzo[k]fluoranthene and between chrysene and triphenylene but can only provide information on benzo[b,j,k]fluoranthene and chrysene/triphenylene as co-elutions. This should be taken into account when evaluating the data.
The available ESB data are summarized in Table S3 (Supplementary material). The table also includes information on additional contaminants that might be of interest for future assessments.
With respect to chemical analyses, the ESB meets all relevant requirements regarding sensitivity, accuracy, repeatability, and reproducibility (Table S2, Supplementary material). Requirements regarding recovery rates are not necessarily fulfilled for the analysis of organic contaminants. Nevertheless, the reliability of data sets concerning the pertained organic contaminants is checked and ensured within each batch of samples by analyzing in-house QA matrix samples, sample material of previous interlaboratory proficiency studies, or certified reference material.
For PAHs, the analytical method used within the ESB-program might lead to an overestimation regarding benzo[b]fluoranthene and chrysene because additional substances are determined (i.e., benzo[j]fluoranthene, benzo[k]fluoranthene, and triphenylene).
D9 assessment based on ESB data
Annual pool samples of blue mussels and eelpout are routinely analyzed for the D9 relevant contaminants. Fairly long time series dating back to the 1990s are available for heavy metals, ∑4 PAHs and B[a]P in mussels and for Pb, Hg, and ndl-PCBs in eelpout (Table S3, Supplementary material). PCDD/Fs + dl-PCBs are currently only determined in eelpout, and time series are relatively short comprising every second year between 2003 and 2015 as well as 2016 and 2017.
With respect to contaminants potentially subject to D9 assessment, TBT, and the WFD priority substances PFOS, PBDE, and HBCDD are included in the present study. For TBT in eelpout and mussels, the available time series end in 2013 (Radermacher 2015). PFOS, PBDE, and HBCDD data have so far been analyzed in archived eelpout samples from every second year between 2003 and 2015 as well as in samples from 2016 and 2017 (Table S3, Supplementary material).
Tables 2, 3, and 4 summarize the contaminant levels in blue mussels and eelpout fillet.
Concentrations in the last year of sampling are given as calculated values (derived from the trend line) and as measured values. In some cases, these values differ significantly (e.g., for TBT by factors of 3.6 and 4.8 in mussels from NS 1 and NS 2, respectively, and for B[a]P by a factor of 3.2 in mussels from BS) indicating that the trend function may not adequately reflect the last year. In these cases, assessing compliance for the last sampling year should consider also the measured concentrations. Note that for ∑4 PAHs in blue mussels from NS 2, TBT in mussels from BS, and HBCDD in eelpout from NS 1 and NS 2, the calculated concentrations are negative and are treated as “not detected.”
North Sea/sampling sites NS 1 (lower Saxony Wadden Sea) and NS 2 (Schleswig-Holstein Wadden Sea)—FAO/ICES division 27.4.b
D9 compliance assessment
Blue mussels from both ESB North Sea sites had Pb, Cd, Hg, ∑4 PAH, and B[a]P levels well below the maximum allowed concentrations in fishery products laid down in Reg. (EC) No. 1881/2006 (amended by Reg. (EC) No. 629/2008, Reg. (EU) No. 1259/2011, Reg. (EU) No. 420/2011, and Reg. (EU) No. 835/2011; EC 2006a, 2008c, 2011a, b, c) (Tables 2, and 3, Figs. 2, and 3). This was still true when the dilution effect caused by breathing water was considered (i.e., multiplying the wet weight concentrations by 3 at NS 1, and 2.4 at NS 2) with the notable exception of ∑4 PAHs in the early assessment years (1985–1995 and 2000 at NS 1, and 1986 and 1990 at NS 2).
Comparing both North Sea sites reveals that concentrations in mussels were always higher at sampling site NS 1 in the Lower Saxony Wadden Sea (Mann-Whitney U test, at least p = 0.02) (Tables 2 and 3, Fig. S1, Supplementary material).
Comparative data are available only for NS 1: The Lower Saxony State Office for Consumer Protection and Food Safety report mean metal concentrations of 0.23 mg kg−1 Pb, 0.11 mg kg−1 Cd, and 0.033 mg kg−1 Hg in blue mussels form the Lower Saxony Wadden Sea in 2016 (LAVES 2017; BMUB 2018a).
The Pb and Cd data fit well to the measured ESB mussel data from 2016 (i.e., 0.17 mg kg−1 Pb and 0.12 mg kg−1 Cd), whereas the Hg concentrations in ESB mussels from NS 1 were lower (i.e., 0.002 mg kg−1). When considering the dilution effect of breathing water, mussels at NS 1 had higher Pb and Cd burdens compared to the LAVES mussels (i.e., 0.52 mg kg−1 Pb and 0.37 mg kg−1 Cd), whereas Hg is still lower (0.006 mg kg−1).
In eelpout fillets, levels of Pb, Hg, PCDD/Fs + dl-PCBs, and ndl-PCBs were below the respective maximum levels allowed in edible fish (EC 2006a, 2008c; EC 2011a) (Figs. 2 and 3).
Since monitoring started in the mid-1990s, levels of Pb and Hg were similar in eelpout from both North Sea sites. Only in 1995 and 2001, higher Pb concentrations were observed at NS 1 (Tables 2 and 3, Figs. 2 and 3, and Fig. S2, Supplementary material).
The eelpout data are in line with the initial assessment of the German North Sea according to Article 8 of the MSFD: Concentrations of PCDD/Fs + dl-PCBs in fillet of cod (Gadus morhua) were below 1 ng kg−1 ww in 2007 (i.e., 0.380 ng kg−1; Karl and Lahrssen-Wiederholt 2009; BMU 2012a) and thus met the threshold value for edible fish. Hg in North Sea fish was also below the maximum allowed concentration (BMU 2012a).
Trend analysis revealed significant decreases (p < 0.01) in metals and PAH contamination of blue mussels from both North Sea sites (Figs. 2 and 3). Decreases were more pronounced in mussels from NS 1 indicating that pollution has declined at this site although it was still higher compared to NS 2 in the last year of sampling (Tables 2 and 3). For PAH, strong decreases were observed at both North Sea sites where PAH levels had been high in the late 1980s and early 1990s. PAH are ubiquitous pollutants in the marine environment originating, e.g., from atmospheric deposition, offshore activities, and operational or accidental spills from ships (Brockmeyer and Theobald 2016). Possibly, the stricter regulations concerning emission and dumping of oil and oily waters that came into force in 1983 (International Convention for the Prevention of Pollution from Ships, MARPOL (1973/1978) and its amendments) resulted in the pronounced decrease of PAHs.
Cd and Hg concentrations in mussels from NS 2 decreased slightly but steadily during the monitoring period. Concentrations between the first and last year differed by only − 23% for Cd and − 19% for Hg. Even though these decreases may not have a direct environmental impact, the steadiness of the decreases nevertheless indicates that pollution is declining in the area.
For eelpout, significant decreases were detected for Hg and ndl-PCBs at NS 1 and for Pb and ndl-PCBs at NS 2 (at least p = 0.01). Pb has also decreased in eelpout from NS 1. This trend, however, was not significant.
In contrast, PCDD/Fs + dl-PCBs increased slightly in eelpout from both North Sea sites (significant trend, p = 0.01, only at NS 2) (Figs. 2 and 3). So far, we have no conclusive explanation for these increases, all the more as no respective trends are observed in sea gull eggs from nearby sites (i.e., island Mellum at NS 1 and island Trischen at NS 2) sampled by the ESB between 2008 and 2015/2016 (data not published yet).
Baltic Sea/sampling site BS (Bodden National Park of Western Pomerania)—FAO/ICES subdivision 27.3d.24
D9 compliance assessment
In mussel samples from the Baltic Sea concentrations of Pb, Cd, Hg, ∑4 PAHs, and B[a]P were well below the maximum levels allowed in fishery products as laid down in Reg. (EC) No. 1881/2006 (amended by Reg. (EC) No. 629/2008, Reg. (EU) No. 1259/2011, Reg. (EU) No. 420/2011, and Reg. (EU) No. 835/2011; EC 2006a, 2008c, 2011a, b, c) (Table 4; Fig. 4). This was, except for ∑4 PAHs in 1993, still true when considering the dilution effect caused by breathing water (i.e., multiplying the wet weight concentrations by 3).
Likewise, compliance with the threshold values was observed for eelpout fillet: During the entire monitoring period, concentrations of the D9 relevant contaminants Pb, Hg, PCDD/Fs + dl-PCBs, and ndl-PCBs in eelpout fillet were below the respective maximum levels allowed in foodstuffs (EC 2006a, 2008b, 2011a) (Table 4, Fig. 4).
These findings correspond to the initial assessment of the German Baltic Sea according to Article 8 of the MSFD: ML compliance was reported for PCDD/F + dl-PCB in fillet of cod (Gadus morhua) and herring (Clupea harengus) from the Baltic Sea (i.e., 0.228–0.631 pg g−1 WHO-TEQ in cod and 2.15–6.28 pg g−1 in herring in 2006; BMU 2012b; Karl et al. 2010) and also for ndl-PCB in herring fillets (14.7–30.4 ng g−1; Karl et al. 2010, ndl-PCB levels in cod were not determined). A separate study analyzed Hg in Baltic Sea fish and found no exceedance of the ML (BMU 2012b).
Similarly, the State Office for Agriculture, Food Safety and Fisheries of Mecklenburg-Western Pomerania reports Pb, Cd, Hg, and PCDD/F + dl-PCB levels in herring fillets from the Western Baltic Sea (ICES Boxes 22 and 24) sampled between 2012 to 2016 that met the respective MLs (BMUB 2018b).
Since monitoring started, concentrations of metals and ∑4 PAHs decreased steadily in blue mussels from the ESB site in the Baltic Sea (p < 0.01). Again, the stricter regulation for ships (e.g., ban of rinsing of tanks and dumping the oily water into the sea, MARPOL (1973/1978) is probably responsible for the decreases. For eelpout, significant decreasing trends (p < 0.01) were detected for Pb, PCDD/Fs + dl-PCBs, and ndl-PCBs, whereas Hg remained more or less constant (Fig. 4).
Evaluation of additional contaminants
Member States may choose to monitor additional contaminants not listed in Reg. (EC) No. 1881/2006 (EC 2006a, 2017a). Potentially problematic contaminants are, e.g., those listed as priority substances under the WFD (EC 2000) and assessed under Descriptor 8-C1 of the MSFD (EC 2017a). For 11 WFD priority substances, wet weight-based EQSs were derived for biota of which nine refer to fish (EC 2013b).
The biota-EQSs represent substance concentrations that are expected to be safe not only for the organisms themselves but also for human consumers (protection goal “human health”) and piscivorous predators (protection goal “secondary poisoning”; EC 2014) because the most sensitive protection goal is decisive for the EQS.
One possible trigger for inclusion in D9 assessment can be the exceedance of the biota-EQS derived for the protection of human health. An inclusion should also be considered if for any of these contaminants increasing trends are detected (Swartenbroux et al. 2010; Zampoukas et al. 2014).
In the present study, four priority substances are exemplarily analyzed that might be of relevance as additional contaminants included in D9 assessment, i.e., PBDE, HBCDD, PFOS, and TBT compounds.
The brominated flame retardants PBDE and HBCDD were widely used, e.g., in electronic products, building material, textiles, and upholstery. Technical Penta-BDE and Octa-BDE mixtures were phased out in the 1990s and are banned in the EU since 2003 (EC 2003a). The Biota-EQS of 0.0085 μg kg−1 fish refers to the sum of the congeners BDE-28, -47, -99, -100, -153, and -154. HBCDD was banned in 2013 (UNEP 2013) but exemptions allowed usage for building insulation materials until 2017. In the years before, it was increasingly used as substitute for PBDE. The biota-EQS for HBCDD is 167 μg kg−1 fish and refers to the sum of α-, β-, and γ- HBCDD.
The fluorosurfactant PFOS was used, e.g., as fabric protector and impregnation agent. Since 2008, PFOS is restricted EU-wide to only a few applications (EC 2006b). Its biota-EQS is 9.1 μg kg−1 fish.
TBT has long been used in antifouling paints on ships and boats. High environmental concentrations were typically associated with marinas and shipyards. In 2003, TBT was finally banned EU-wide (EC 2003b). No biota-EQS exists for TBT, but OSPAR and HELCOM have set an environmental assessment criterion (EAC) of 12 μg kg−1 dry weight for TBT in bivalves that relates to toxic effects on bivalves and the protection goal secondary poisoning (OSPAR 2004). Beyer et al. (2017) converted the original OSPAR EAC value to a wet weight-based EAC of 2 μg kg−1 ww for formulation-based TBT (using an average dry mass content for mussels of 17.38%). Blue mussels from the ESB sites had lower average dry mass contents, and accordingly, the respective wet weight-based EACs are lower (i.e., 1.13 μg kg−1 ww and 0.88 μg kg−1 ww for mussels from NS 1 and NS 2, respectively, and 0.63 μg kg−1 ww for mussels from BS; Tables 2, 3, and 4). Norway has derived a biota quality standard of 150 μg kg−1 ww for TBT compounds (NEA 2016).
Based on retrospectively analyzed ESB samples of eelpout and blue mussel, temporal trends for TBT, PFOS, PBDE, and HBCDD have been calculated (Table S4, Fig. S4, S5, Supplementary material). Figure 5 summarizes the data from all three sampling sites.
The retrospective analysis revealed high TBT concentrations in blue mussels from the Lower Saxony Wadden Sea (NS 1) in the 1980s and 1990s that exceeded the EAC value by far (factor up to 16 based on dry weight concentrations) (Fig. 5a). Since the EU-wide ban in 2003, concentrations have declined significantly (p < 0.01, Table S4, Fig. S4, Supplementary material) and complied with the EAC of 12 μg kg−1 dw in 2010–2013. For NS 2 and BS TBT, data are available since 2004 and 2005, respectively. Lowest contamination was detected at NS 2 where mussels are not directly exposed to emissions from marinas or shipping traffic. Similar to NS 1 TBT, levels have decreased significantly in mussels from BS and NS 2 (p < 0.01, Table S4, Fig. S4 and S5, Supplementary material) and met the EAC in 2008 (NS 2) and 2011 (BS). Time series for TBT in mussels end in 2013 because new increases are not expected (for more details see Rüdel et al. 2003, 2009 and Radermacher 2015).
TBT levels in eelpout fillets were mostly higher than in mussels (Tables 2, 3, and 4, Fig. S3, Supplementary material). This is at least partly due to different exposure conditions as sampling sites for mussels and eelpout are not identical.
PFOS, PBDE, and HBCDD were analyzed only in eelpout because their EQSs refer to fish.
PFOS concentrations in eelpout fillet were similar at all three ESB sites and well below the respective EQS of 9.1 μg kg−1 (Fig. 5b). Significant changes since 2003 were only detected in fish from NS 2 (decreasing trend, p = 0.01; Table S4, Fig. S4, Supplementary material). For PFOS, the critical protection goal behind the EQS is human health. Accordingly, the results indicate that PFOS in marine fish from the three coastal sampling sites pose no risk to human consumers.
The picture is different for PBDE with 100% EQS exceedance at all three sites and in all years (Fig. 5c). Despite significant decreases (p < 0.01, Table S4, Fig. S4 and S5, Supplementary material), PBDE concentrations in eelpout fillets from all three sites still exceeded the EQS in 2016 and 2017 by factors of 4–10. The critical protection goal behind the EQS is human health implying that, despite decreasing trends, PBDE in marine fish from the three coastal sites can still pose a risk for human consumers.
HBCDD concentrations in eelpout fillet were always way below the EQS of 167 μg kg−1 (Fig. 5d). In most years, concentrations ranged between 0.11 and 1.5 μg kg−1 with higher levels detected only in 3 years (where sample handling or measurement errors cannot be excluded). Significant decreases were detected at NS 1 (p = 0.03) and BS (p < 0.01; Table S4, Fig. S4 and S5, Supplementary material). The EQS for HBCDD is based on the protection goal secondary poisoning. The quality standard derived for human health is about 36 times higher (i.e., 6100 μg kg−1, EC 2014). According to these data, HBCDD in marine fish from all three ESB sites pose no risk for human consumers.