2.1 Study species and area

Fucus distichus and Fucus spiralis (also known as rockweed or bladderwrack) are commonly found in the intertidal zone of rocky shores of the Salish Sea, a large (approximately 18,000 km²) international inland sea whose watersheds drain a wide range of land-use types including highly urbanized areas, residential developments, agricultural areas, and relatively pristine habitats. Harvesters in Washington (WA) State and in British Columbia (BC), gather both vegetative blades and reproductive receptacles from Fucus for use in foods, teas, and medicines [42]. Bull or bull whip kelp Nereocystis luetkeana is a subtidal brown alga that is common in nearshore habitats. Both blades and stipes are harvested for food, medicine, agriculture, and mariculture [42, 43].

Intertidal Fucus distichus from 38 sites, intertidal Fucus spiralis from three sites, and subtidal Nereocystis luetkeana from 18 sites were collected from the WA State (US), and BC (Canada) coasts (Fig 1) of the Salish Sea; sixteen sites had both F. distichus and N. luetkeana. Sites represented a range of potential contamination. Samples were collected from sites known to be or suspected of being contaminated based on current or historic activities, including mines (e.g., Britannia Mine), smelters (e.g., Ruston Way in the Asarco’s Tacoma smelter plume), coal gasification plants and refineries (e.g., Rock Bay, Fidalgo Bay), ports, shipyards, and ferry terminals (e.g., Duke Point, Point Hope, Mukilteo Ferry, Tsawwassen Jetty, Smith Cove, Port Angeles, Victoria Jetty, Port of Everett), and present and former pulp and paper mills (e.g., Crofton Mills, Elk Falls, and Goodridge Mill). They also included sites at or near National Oceanic and Atmospheric Administration (NOAA) Mussel Watch or Washington Department of Fish and Wildlife (WFDW) transplanted (i.e., caged) mussel sites. These sites included Smith Cove, Mukilteo Ferry, Wing Point, Waterman Point, Four Mile Rock, Post Point, Port of Everett, and Ruston Way [35, 44, 45]. Additional sites included areas where seafood was historically or is currently gathered by Indigenous peoples on Canadian First Nation reserve lands, on Washington tribal lands, in “usual and accustomed” harvesting areas, and in “traditional territory” with minimal history of upland industrial development.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Sites sampled in the Salish Sea.

AB: Agate Beach, AS: Ambleside, BB: Botanical Beach, BM: Britannia Mine, BR: Brothers Islands/Duntze Head, CA: Cape Mudge, CB: Coles Bay, CH: Camano Head, CI: Chatham Islands/Tl’chés, CM: Crofton Mill, DB: Deep Bay, DP: Duke Point Industrial Park, EF: Elk Falls Pulp Mill, CR: Campbell River, EH: Esquimalt Harbour, ES: East Sound, FW: Fresh Water Bay, FB: Fidalgo Bay Aquatic Reserve, FM: Four-Mile Rock, FR: Fort Rod Hill, GM: Goodridge Mill, PO: Portage Inlet HB: Hagan Bight, JB: Jefferson Beach, KB: Kulleet Bay, LT: Lone Tree Point, MF: Mukilteo Ferry, PA: Port Angeles, PE: Port of Everett, PH: Point Hope Shipyard, PI: Penelakut Island, PP: Post Point, PO: Portage Inlet, RB: Rock Bay, RW: Ruston Way, SB: Sooke Bay, SC: Smith Cove, SI: Senanus Island, SN: Sansum Narrows, SQ: Squaxin Island, TJ: Tsawwassen Jetty, VI: Vashon Island, VJ: Victoria Jetty, WP: Waterman Point, WG: Wing Point. Red circles indicate sites where only F. distichus were sampled, blue circles indicate sites where only F. spiralis were sampled, green squares indicate sites where only N. luetkeana were sampled, and green triangles indicate sites where both F. distichus and N. luetkeana were sampled.

https://doi.org/10.1371/journal.pone.0269269.g001

2.2 Sample collection and preparation

Sampling occurred from June to September 2015, during the growing season, with sustainable harvesting methods that followed the laws for recreational shellfish and seaweed license holders issued by WDFW and the 2015 regulations issued by the Aquatic Plants Program, Ministry of Agriculture, BC, Canada. All sites were sampled one time, except Vashon Island, which was sampled in June, July, and August. Beaches were divided into left, right, and center segments, and samples were collected from each segment. When relevant, field site access was approved by duly authorized First Nation and Indian Tribal Community staff. Permits were not obtained or required due to the specific nature of the work. Fidalgo Bay Aquatic Reserve seaweed site access was provided by Birdie Davenport. Botanical Beach Provincial Park seaweed site access was provided by Kristine Pearson, Pacheedaht First Nation. The Canadian Forces Base at Esquimalt Harbour provided access to seaweed sample sites in Esquimalt Harbour area. All other sites were publicly accessible or had permission as described above. Geographic or GPS (global positioning system) coordinates were recorded; however, we are not publishing the coordinates because of the sensitive nature of providing specific locations on Indian Reserve Lands or seaweed sites in First Nation or Tribe Territories.

N. luetkeana non-reproductive and reproductive blades were cut from individuals in the shallow subtidal zone, blades were removed at least 70 cm above the bulb with a ceramic-bladed knife. F. distichus and F. spiralis thalli, including blades, midribs, and reproductive and non-reproductive apical tips, were collected from the intertidal zone with a ceramic-bladed knife so that no more than half of any individual was removed. F. spiralis was only sampled when F. distichus was not present on a beach.

Seaweeds were rinsed with seawater at the sampling site in cotton mesh bags. The bags containing the seaweed were rinsed 12 times in seawater from the site, drained, and immediately transferred and stored in polyethylene (Ziploc^™) bags in a cooler with ice packs. Within two days, they were transferred to a -80°C ultra-low freezer, except for three remote sampling locations where samples were on ice packs for up to four days. To make a single composite sample, seaweeds were thawed, and visible fauna and epiphytes were removed. Seaweeds were not rinsed before processing as is consistent with the practices of many consumers [46]. For each site, to create the composite sample, equal portions of samples from the left, right and middle segments of the beach/kelp bed were processed together by hand-chopping the mixture into a slurry with ceramic-blade knives. Between samples, equipment was cleaned with Tergazyme^® detergent and de-ionized water. Homogenized seaweed (approximately 20 g per composite sample per set of analyses) was transferred to certified-contaminant-free, amber glass jars with plastic lids, and stored at -80°C until it was shipped on ice 1–3 weeks later for chemical analyses.

2.3 Chemical analysis

A complete list of analytes and measured concentrations along with lower limits of quantitation (LOQs) are given in [47]. F. spiralis and F. distichus tissues (but not N. luetkeana due to financial constraints) were analyzed for 22 organochlorine pesticides, 40 PCBs, 15 PBDEs and 51 polycyclic aromatic hydrocarbons (PAHs) by the NOAA / Northwest Fisheries Science Center Environmental Chemistry Lab (Seattle, Washington, USA) using the gas chromatography/mass spectrometry (GC/MS) method and quality assurance plan of Sloan et al., [48, 49]. The mean concentrations of analytes in method blanks were below detection limits or at least ten times lower than the lowest field sample for which the measured concentration was above the detection limit [47]. F. spiralis and F. distichus tissues were also analyzed for the 15 PCDD/Fs with the WHO 2005 TEFs [50] and the dioxin and furan octa homologues by AXYS Analytical Labs (Sidney, BC, Canada; currently SGS AXYS) using USEPA Method 1613b in accordance with internal QA/QC policies and procedures. To measure concentrations of metals, F. spiralis, F. distichus, and N. luetkeana tissues were dried in a 95°C oven overnight and acid digested with trace metal grade HNO₃ and HCl. Percent moisture was determined prior to the digestions. The digested samples were analyzed at the University of California (Davis, California, USA) for 17 metals (beryllium [Be], vanadium [V], chromium [Cr], manganese [Mn], iron [Fe], cobalt [Co], nickel [Ni], copper [Cu], zinc [Zn], As, selenium [Se], molybdenum [Mo], Cd, barium [Ba], Hg, thallium [Tl], Pb) using an Agilent 7700 ICP-MS in He collision mode. The LOD and LOQ were determined according to FDA Elemental Analysis guidelines for food and related products.

A variety of methods (e.g., [51]) have been proposed to handle values that fall below the lower limit of quantitation (LOQ) for the purpose of data analysis. These range from a highly conservative approach of setting these values to the LOQ to a much less conservative approach, from a human health perspective, of setting them to 0. Unless otherwise noted, we chose a moderately conservative approach and used ½ the LOQ for our analyses. Thus, if a sample was below the LOQ, one-half the LOQ was used for summed concentrations of the congeners we analyzed (e.g., Σ₄₀PCBs, ΣPBDEs, ΣPCDD/F TEQs), calculations of means, standard deviations, and medians, and for statistical analyses, unless otherwise stated. TEQs were calculated with 2005 WHO TEFs from [50]. Methods for calculating total chlordanes, total DDTs, Σ₄₀PCBs, total chrysenes, ΣPBDEs, and ΣPCDD/F TEQs are included in the S1 Text.

2.4 Screening level calculations

To assess potential health risks to consumers, we compared seaweed contaminant concentrations to calculated health-based screening levels for the 22 persistent organic pollutants (POPs), seven PAHs (Table 2), and eleven metals (Table 3) for which RfDs or oral CSFs exist. We calculated two screening levels based on a consumption rate of 5 g dry weight (DW) of seaweed per day, which is the typical amount consumed in a meal [52]. This rate is comparable to values reported for South Korea at 8.5 g/adult/day [53], India and China at 5.2 g/capita/day [54–56], and Japan at 4 to 10.4 g/adult/day [57, 58].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Summary of screening levels and analytical results for organic contaminants.

https://doi.org/10.1371/journal.pone.0269269.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Summary of screening levels and analytical results for metals.

https://doi.org/10.1371/journal.pone.0269269.t003

The RfD-based screening level (SL_RfD) for adults was based on RfDs (mg/kg body weight/day) obtained from the Integrated Risk Information System (IRIS) database provided by the USEPA and was calculated as: (1) where SL_RfD (in mg chemical/kg DW of seaweed) is the reference dose-based screening level and is used for non-carcinogenic effects, BW is body weight (assumed to be 80 kg), CR is the consumption rate (g DW of seaweed/day), and 1,000 is a unit correction. We did not speciate metals in our chemical analysis, however, in the IRIS database there are oral RfDs for five metal species (Cr (III), Cr (IV), methyl Hg, inorganic As, and V pentoxide). To be conservative, we used the RfD for the speciated forms of these metals to calculate the screening levels; for Cr, we used Cr (IV). The second screening level (Eq 2) was based on oral CSFs (1/mg/kg-day) for adults obtained from the USEPA IRIS database. (2) where SL_CSF (in mg chemical/kg DW seaweed) is the CSF-based screening level and is used for carcinogenic effects, and ARL is the acceptable risk level of 1E-6. The upper-bound slope factor in the higher risk and persistence tier was used for PCBs. For convenience, when making comparisons to organic contaminants, screening levels were converted to μg chemical/kg seaweed.

We then compared the seaweed contaminant concentrations to the calculated screening levels. To guide acceptable site-specific consumption rates for sites where contaminant concentrations exceeded screening levels, a new maximum consumption rate was calculated by solving for the consumption rate in Eqs 1 and 2 with the screening level set at the concentration (in mg/kg DW) that we measured in the seaweed. If more than one analyte exceeded the screening level at a site, the most conservative new maximum consumption rate was identified. In the case of Pb, which does not have an assigned RfD or CSF, we compared our seaweed concentrations to the lowest international limit (China: 1 mg/kg DW) and calculated new consumption rates with that value.

2.5 Total diet study

To provide context for the measured levels of focal contaminants in F. distichus and N. luetkeana, we calculated the amounts of contaminants that would be consumed in 5 g DW portions of seaweed and in established portion sizes for commonly consumed foods based on contaminant concentrations and portion sizes in [59–62]. We did not compare F. spiralis to commonly consumed foods due to limited samples (n = 3). To identify locations where a contaminant in seaweed is elevated compared to the mean, we flagged any samples where concentrations/portion exceeded twice the standard deviation for that species. These samples were considered statistical outliers and with only a few exceptions (e.g., Deep Bay, Vashon Island in June, and Jefferson Beach), they were collected near highly urbanized areas. The use of these statistical outliers was only for the total diet study comparisons. For each seaweed species and focal contaminant, we compared the concentration per portion in commonly consumed foods with the concentrations per portion of seaweed for samples contaminated enough to be identified as statistical outliers and with the averaged concentrations per portion of seaweed for all other remaining sites.

2.6 Interspecific comparisons

Comparisons of select metal concentrations in F. distichus and N. luetkeana were made with paired t-tests after testing for normality. The Bonferroni-corrected α value of 0.017 was used to account for making multiple comparisons.

2.7 Inclusivity in global research

Additional information regarding the ethical, cultural, and scientific considerations specific to inclusivity in global research is included in the S1 Checklist.

3.1 Detection of organic compounds

Although some contaminants were at high enough concentrations to result in recalculated consumption rates, concentrations of organic contaminants in Fucus were generally low, with most samples below the analytes’ LOQs (Table 2 and [47]). For analyses of POPs and PAHs, LOQs for compounds that were below the LOQs ranged from 0.139 to 3.12 μg/kg; for PCDD/DFs, they ranged from 0.089 to 0.508 μg/kg. Fifteen percent of the pesticide measurements were above the LOQ (3311 total measurements; [47]). Of the pesticides that had screening levels (Table 2), concentrations of only four (α-HCH, γ-HCH, 4,4’-DDE, and 4,4’ DDT) were above LOQs in F. distichus, and this only occurred at Rock Bay. For F. spiralis, they were only above LOQs for 4,4’-DDD, and 4,4’-DDE at Portage Inlet. Only 26 percent of individual PCB congener measurements at all sites exceeded LOQs (1720 total measurements; [47]). In F. distichus, concentrations of 12 of the 40 PCB congeners were below LOQs at all sites, whereas in F. spiralis, 16 were below LOQs at all sites (Fig 2a). Only 35% of the PCDD/DF congener measurements at all sites exceeded LOQs (2884 total measurements; [47]). Most measurements of 2,3,7,8-TCDD, the only dioxin or furan congener with a screening level, were below LOQs, with only F. distichus at Ambleside and Deep Bay exceeding them (Table 2). Four of the PBDEs congeners were not able to be identified. Of the 11 identified congeners (PBDEs 28, 47, 49, 66, 85, 99, 100, 153, 154, 155, and 183), only two with screening levels (PBDE 47 and 99) exceeded their LOQs at one or more sites. PBDE 47 and 99 were above the LOQ at Rock Bay in F. distichus and PBDE 49 was above LOQs in 43% of the F. distichus samples and 66% of the F. spiralis samples. Similarly, concentrations of the PAHs with screening levels were below LOQs at most sites, except naphthalene, for which all samples were above LOQs (Table 2). Acenaphthene, anthracene, and fluorene concentrations were above LOQs in F. distichus at two sites in Victoria Harbour (Point Hope and Rock Bay), and anthracene and fluorene concentrations were above LOQs in F. spiralis at Portage Inlet, which is also in Victoria Harbour. Fluoranthene concentrations were above LOQs at the three inner Victoria Harbour sites (F. spiralis at Portage Inlet, and F. distichus at Rock Bay and Point Hope), and also in F. spiralis at Fidalgo Bay, and in F. distichus at Crofton Mill and Port Angeles. BaP concentrations exceeded LOQs in Fucus spp. at the three Victoria Harbour sites and in F. spiralis at Fidalgo Bay (Fig 2b).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2.

Concentrations of focal contaminants in Fucus spp. and Nereocystis at Salish Sea sites from the southernmost sites (left) to the northernmost sites (right). Bars are concentrations of a) Σ₄₀PCBs (μg^.kg^-1 DW), b) benzo[a]-pyrene (BaP, μg^.kg^-1DW), c) total arsenic (mg^.kg^-1 DW), d) cadmium (mg^.kg^-1 DW), e) lead (mg^.kg^-1 DW), and f) total mercury (mg^.kg^-1 DW) calculated using the limit of quantification (LOQ)/2 when measured values were less than the LOQ. Upper ends of the vertical lines in a) are the calculated Σ₄₀PCB concentrations when the LOQ was used for concentrations that were less than the LOQ. Lower ends of the bars are calculated values obtained when zero was used for the values less the LOQ. For a) Σ₄₀PCBs and b) benzo[a]-pyrene, black dashed horizontal lines indicate the cancer slope factor-based screening levels (SL_CSF). For c) total As, d) Cd, e) Pb, and f) total mercury, the upper black dashed horizontal line is the highest international limit (see Table 1) for the contaminant, and the lower black dashed horizontal line, if present, indicates the lowest international limit. Red and orange bars are concentrations in Fucus distichus that exceed or are less than the LOQ, respectively; yellow hatched bars or white hatched bars are concentrations in Fucus spiralis that exceed or are less than the LOQ, respectively; green cross-hatched or blue cross-hatched bars are concentrations in Nereocystis luetkeana that exceed or are less than the LOQ, respectively. Site codes are as in Fig 1.

https://doi.org/10.1371/journal.pone.0269269.g002

3.2 Screening levels for organic chemicals

Except for Σ₄₀PCBs and BaP, concentrations of all organic contaminants were below the RfD and CSF-based screening levels at all sites (Table 2). BaP concentrations only exceeded the CSF-based screening level of 16 ng/g^-1 dry weight (Fig 2b) at two of the 41 sites: Portage Inlet and Rock Bay. None of the Σ₄₀PCBs concentrations exceeded the RfD-based screening level regardless of how congener concentrations below the LOQ were treated. The number of sites with concentrations of Σ₄₀PCBs that exceeded the CSF-based screening level, however, depended on how concentrations below the LOQ were treated (Table 2). When Σ₄₀PCBs were calculated using 0 for congener concentrations less than the LOQ, 18% of F. distichus and 33% of F. spiralis samples had concentrations that exceeded the CSF-based screening level (Fig 2a). When they were calculated using ½ the LOQ for values less than the LOQ, 70% of F. distichus and 67% of F. spiralis samples had concentrations that exceeded the CSF-based screening level. Finally, when Σ₄₀PCBs were calculated with values less than the LOQ set to the LOQ, 98% of F. distichus and all F. spiralis concentrations exceeded the CSF-based screening level. The timing of sampling could also affect whether screening levels were exceeded as Σ₄₀PCBs concentrations in F. distichus calculated with values below the LOQ set to zero at Vashon Island were above the CSF-based screening level in June (VI6 in Fig 2a), but below it in July (VI7) and August (VI8); however, further work is needed to confirm a temporal pattern because these measurements were not replicated and were only made at one site.

Other organic contaminants in the seaweed samples were well below screening levels. For example, organochlorine pesticide concentrations at Rock Bay and Portage Inlet, which were among the highest concentrations we measured, were only 0.4–27% of the CSF-based screening levels or 0.005–0.04% of the RfD-based screening levels (Table 2). Only two samples had detectable 2,3,7,8-TCDD concentrations, but those concentrations were less than 0.1% of the screening level. Likewise, the ΣPCDD/F TEQs were well below the RfD-based screening level at all sites. Concentrations of ΣPBDEs and all PBDE congeners for which RfDs exist (PBDE 47, 99, and 153) were always less than 0.1% of those screening levels for F. distichus and F. spiralis at all sites (Table 1). Furthermore, maximum concentrations of acenaphthene, anthracene, fluorene, and naphthalene also did not exceed 0.002% of the RfD-based screening levels. The highest concentration of total chrysenes was 69% of the CSF-based screening level, and the highest fluoranthene concentration was 0.013% of the RfD-based screening level (Table 1).

3.3 Detection of metals

Concentrations of many metals (Ba, Cd, Co, Cr, Fe, Mn, Mo, total As, V and Zn) were above LOQs at most sites. However, Cu, Pb, and Se concentrations were generally below LOQs. For all metal analyses with a measured value below the LOQ, the LOQ was 0.1 mg/L [47]. Concentrations of Be and Tl in all three species were always below LOQs. Ni concentrations were above the LOQ in 65% and 100% of the F. distichus and F. spiralis, respectively, and none of the N. luetkeana. Conversely, total Hg was above the LOQ in 78% of the N. luetkeana compared to 18 and 0% of the F. distichus and F. spiralis, respectively (Table 3).

3.4 Screening levels and international limits for focal metals

An RfD or CSF for Pb has not been assigned, so comparisons to the calculated screening levels were not conducted. Except for As, the remaining metal concentrations never exceeded US RfD-based screening levels at any site (Table 3). Total As concentrations in all samples exceeded the US RfD- and CSF-based screening levels for inorganic As, as well as the recommended levels for inorganic As in China, France, and Australia/New Zealand (Fig 2c; Table 1). All Cd concentrations were above the levels recommended by Australia/New Zealand, France, and China (Fig 2d, Table 1). However, concentrations of Cd in all three species never exceeded the European Commission maximum level of 3 mg^.kg^-1 wet weight, assuming an 85% moisture content for the N. luetkeana and using actual wet:dry weight ratios for Fucus spp. Chinese and French recommended levels for Pb were exceeded only in Point Hope and Rock Bay for F. distichus and Portage Inlet for F. spiralis but were never exceeded in N. luetkeana samples (Fig 2e, Table 1). Total Hg did not exceed the screening level calculated with the RfD for methyl Hg in any samples (Fig 2f, Table 3), however, detectable total Hg concentrations exceeded the French recommended level of 0.1 mg^.kg^-1 DW for Hg in 15% of F. distichus samples and 78% of N. luetkeana samples. They did not exceed the French recommended level for Hg in any F. spiralis samples (Fig 2f, Table 1).

3.5 Interspecific, temporal, and latitudinal differences in metal concentrations

Concentrations of total As, Cd, and total Hg were 184, 111, and 151% higher, respectively, in N. luetkeana than in F. distichus (Fig 3) when they were sampled from the same sites at the same time (N = 15 sites, paired t-tests: total As t = 16.25, P<0.001; Cd t = 10.13, P<0.001; total Hg t = 6.34, P<0.001). We saw evidence of temporal variability in total As and Cd concentrations in both F. distichus and N. luetkeana at Vashon Island (Fig 3), which were sampled in June, July and August; however, because these were single composite samples, it is not possible to determine whether these differences were significant. Although concentrations of many contaminants differed among sites, there were no obvious latitudinal patterns in their abundances (Fig 2). Correlations between latitude and total As, Cd, total Hg, Pb, BaP and Σ₄₀PCBs in F. distichus were never statistically significant (|Spearman’s rho|<0.293, P>0.130). No statistically significant correlations existed between latitude and total As, Cd, total Hg, and Pb in N. luetkeana (|Spearman’s rho|<0.281, P>0.431). Likewise, collection dates were not significantly correlated with total As, Cd, total Hg, Pb, BaP and Σ₄₀PCBs in F. distichus (|Spearman’s rho|<0.188, P>0.339), nor with total As, Cd, total Hg, and Pb in N. luetkeana (|Spearman’s rho|<0.804, P>0.090).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Concentrations (in mg^.kg^-1 DW) of a) total arsenic (As), b) cadmium (Cd), and c) total mercury (Hg) in Fucus distichus (FD) and Nereocystis luetkeana (NL) at sites where both species were collected (N = 16).

Data are means ± 1 SD. Concentrations are significantly (P<0.001) higher in N. luetkeana than F. distichus for all three metals (paired t-tests).

https://doi.org/10.1371/journal.pone.0269269.g003

4.1 Contaminant concentrations

Many chemicals that can biomagnify were measured at low concentrations in our seaweed samples possibly because of the seaweeds’ annual (Nereocystis) or biennial (Fucus spp.) life cycles, low lipid content, and position at the bottom of the food web. However, PCBs were detected in all our Fucus and Nereocystis samples, indicating that they are ubiquitous in edible seaweeds throughout the Salish Sea. PCBs are of high concern in the Salish Sea because they occur in a wide range of species at concentrations high enough to cause concern for both the marine organisms and humans who consume them. When our Σ₄₀PCB concentrations for F. distichus were compared to the Venice Lagoon (Italy), the only location for which comparable data for Fucus spp. could be found, those in the Salish Sea were higher (S1 Table).

Even if our Σ₄₀PCBs are calculated using 0 for congeners below LOQs, all sites in the current study had higher concentrations than the average of 3 μg/kg DW reported for an unidentified species of Fucus [64] (likely F. virsoides based on distribution information from [65]). BaP, the other focal organic contaminant, was only detected in seaweeds collected at 5 sites in the Salish Sea, and generally fell within the ranges reported for Fucus spp. elsewhere, although the Salish Sea site with the highest BaP concentrations, Rock Bay, had a concentration (64.6 μg/kg DW) slightly higher than the maximum value found in F. vesiculosus (64 μg/kg) from Norway [66]. It should be noted that comparisons to literature-reported seaweed concentrations come with caveats. Sites from these other studies may have been selected to represent extreme levels of contamination, different analytical techniques may have been used, data analysis is not consistent, interspecific variation in uptake rates may occur, and there are limited studies for some of the contaminants we investigated [67]. A comparison is useful, however, as a way of generally assessing whether Salish Sea seaweeds demonstrate an extreme level of contamination. Details on how literature was selected for comparison are in the S2 Text.

In contrast to the organic chemicals, many metals were above LOQs at most of our sites. Higher concentrations of metals are predicted as seaweeds have been shown to hyper-accumulate metals [68–70]. Brown seaweeds are highly effective at bioabsorbing and bioaccumulating metals due to their outer cell walls and matrix structures, which have abundant mucilaginous polysaccharides such as alginate and fucoidans and metal-binding metallothionein proteins and phlorotannins [69, 71]. Like PCBs, the focal contaminants Cd and total As were detected in all our Fucus and Nereocystis samples, supporting that they too are widely distributed in these Salish Sea seaweeds. However, total Hg and Pb were not commonly detected, with only 18 and 25% of samples above the limits of quantification.

The average concentrations of Cd, Pb, and total As in F. distichus and F. spiralis (Table 2) were within ranges of these metals measured in Fucus spp. in other studies (S2 Table). Likewise, Pb and total As in N. luetkeana are comparable to those in previous studies (S2 Table). At the sites where Hg was detectable (Fig 2f), our concentrations of total Hg in Fucus were higher than previously reported (S1 Table), except for a single report for F. vesiculosus in Norway [72], which had total Hg concentrations that exceeded our highest measurement by more than fourfold. Average total Hg concentrations in N. luetkeana (0.12 mg/kg) were also higher than in earlier reports from British Columbia where Hg was below detection levels [72, 73]. Average Cd concentrations (5.69 mg/kg) in N. luetkeana (Table 3) were 2 to 19 times greater than previously published concentrations (S2 Table) from British Columbia [72, 73].

Some of these differences may be attributed to our study design, which included sites ranging from those with minimal history of upland industrial development to sites known to be contaminated. The differences may also be the result of the Salish Sea’s hydrology and geological history, which causes background concentrations of many elements, such as As and Cd, to be elevated [74, 75], and affects their concentrations in marine biota. For example, clams, crabs, and finfish in Puget Sound generally have high levels of inorganic As, likely from natural sources, such as groundwater runoff from areas with volcanic activity [76]. Likewise, high Cd concentrations in mussels from the Salish Sea and other sites off the coast of Washington have been attributed to the upwelling of Cd-rich waters [77]. However, natural sources are not the only concern as sediment core profiling at two sites in Puget Sound has shown that many elements, such as As, Pb and Cu, have increased as a result of urbanization of the surrounding watershed [78]. Naturally occurring metals, therefore, are expected to contribute to a baseline concentration of metals in the seaweed with elevated levels near urbanized and industrial areas.

It should be noted that average concentrations of focal contaminants were sometimes skewed by a few sites that had high concentrations, which were identified as statistical outliers for our diet comparisons (Fig 4). These included Σ₄₀PCB levels in F. distichus at Rock Bay, Vashon Island (June only), and Smith Cove; BaP in F. distichus at Rock Bay; total As in F. distichus at Vashon Island (August only); total As in N. luetkeana at Brothers Islands/Duntze Head; Cd in F. distichus at Jefferson Beach; Pb in F. distichus at Point Hope and Rock Bay; Pb in N. luetkeana at Four Mile Rock and Victoria Jetty; total Hg in F. distichus at Deep Bay and Waterman Point; and, total Hg in N. luetkeana at Victoria Jetty. All sites except Deep Bay (DB) on central Vancouver Island were in South Puget Sound near Seattle (US) or in Victoria/Esquilmalt Harbour (Canada). These areas have a history of heavy industrialization, are highly urbanized with high amounts of impervious surfaces, and are influenced by large drainage basins.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Comparisons of a) Σ₄₀PCBs, b) benzo[a]-pyrene, c) total arsenic, d) cadmium, e) lead, and f) mercury in F. distichus (red bars) and N. luetkeana (green bars) to commonly consumed foods.

Sites for which contaminant concentrations exceeded the mean + 2SD are shown separately. Bars labeled as F. distichus or N. luetkeana are mean concentrations (+ 1SD) from all sites that do not exceed the mean + 2SD. *Means that excluded sites with values greater than the mean + 2SD. Data for commonly consumed foods were obtained from [59–62]. More information about commonly consumed foods and the sources for this data are provided in S3 Table.

https://doi.org/10.1371/journal.pone.0269269.g004

As a result of historic activities, the region includes several known or suspected contaminated sites. An example in the United States Salish Sea is the Lower Duwamish Waterway (LDW) in South Puget Sound which is downstream of the Green River Basin and flows into Elliott Bay. A five-mile stretch of this waterway is a Federal USEPA Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA) Site with PCBs, carcinogenic PAHs, and Hg in sediments at concentrations that exceed acceptable levels for human health [79]. Seventeen formal Washington State cleanup Sites under the Model Toxics Control Act, the state environmental cleanup law, are part of the 20,400-acre LDW source control area [80]. In Canada, the lower portion of Victoria Harbour was a shipping and industrial port but has evolved to serve recreational boating and mooring, commercial and light industry, and residential needs as well. Portage Inlet, in the upper Harbour, was not heavily industrialized because of limited navigable waters, but is surrounded by residential areas and used by small pleasure craft. Industrial contributions to contaminants in Esquimalt Harbour have included sawmills, and the presence of Royal Navy and ship servicing activities [81, 82].

Although historic and current industrial activities are well-established sources of persistent environmental contaminants such as PCBs and metals, there is a growing recognition that nonpoint source pollution is a major and current contributor to marine contamination. For example, surface runoff, which includes stormwater, nonpoint source overland flow and groundwater discharge to surface waters, is the largest contributor of PCBs, PAHs, pesticides and metals to Puget Sound [83, 84]. Contaminant mass loading is also affected by drainage basin size and types of land use. Large areas of forested lands contribute high volumes of surface runoff but low concentrations of contaminants, whereas small urbanized and agricultural areas contribute smaller volumes but higher contaminant concentrations [85]. Likewise, modeling and empirical studies of Victoria (Canada) indicate that stormwater runoff (defined here to include surface runoff, precipitation, snowmelt and groundwater that is hydrologically connected to stormwater drainage systems) could be the primary route for surface water contaminants including metals and organic chemicals, such as PCBs, in Victoria Harbour [86, 87] and could contribute to high contaminant concentrations in seaweeds near Victoria.

4.2 Dietary implications

In this study, we calculated the amounts of F. distichus and N. luetkeana from individual sites that could be consumed without exceeding the following limits: CSF-based screening levels for Σ₄₀PCBs, BaP, and RfD-based screening levels for Cd or methyl Hg from the risk-based USEPA recommendations; or international limits in seaweeds for Pb. We did not complete these calculations for As given the uncertainties in the amount of inorganic As in the samples. In our analysis, we compared five of the focal contaminant concentrations in seaweeds to RfD and CSF-based screening levels, where the CSF is specifically related to cancer risk. US regulations use 1 x 10⁻⁴ to 1 x 10⁻⁶ for an acceptable excess lifetime cancer risk, with the latter being the most protective. The choice of the risk level to use is a policy decision as opposed to a scientific decision. The CSFs we used are for an acceptable cancer risk at 1 x 10⁻⁶ but are based on a standard methodology that likely overestimates cancer risk by several orders of magnitude. This results in an estimate of risk that is more protective than the 1 x 10⁻⁶ cancer risk [88]. Additionally, there are different levels of confidence for whether a chemical is a human carcinogen. Contaminants are classified into categories based on the confidence that they cause cancer in humans proceeding from high to low evidence levels: “known human carcinogens”, “probable human carcinogens”, and “possible human carcinogens” [89]; these classifications were updated with new descriptors such as “carcinogenic to humans” to replace “known human carcinogens” [90]. Finally, there is concern that unintended consequences, such as avoiding a food with potential health benefits, may result from perceived concerns about potential cancer risk. Many of these considerations for fish consumption are appropriate for seaweed, which present consumers with a trade-off decision between potential health benefits and cultural well-being, and potential health risks from contaminant exposure.

Of all contaminants for which we could calculate screening levels based on USEPA information, only Σ₄₀PCBs and BaP had samples with exceedances. This is in agreement with a number of studies that have identified PCBs and high molecular weight PAHs as contaminants of high concern in the Salish Sea, regarding both organismal and human health. PCBs have long been monitored in a number of Puget Sound indicator species of ecosystem health including English sole, a benthic flatfish (Parophrys vetulus), Pacific herring, a pelagic planktivore (Clupea pallasii–[36], bay mussels, a nearshore bivalve (Mytilus trossulus–[35, 45]), and Chinook salmon (Oncorhynchus tshawytscha) in both juvenile [91] and adult [92] forms. The Washington State Department of Health recommends meal limits on salmon and other Puget Sound species because of high PCB levels [93], and PCBs are a primary target for reductions in Puget Sound to restore ecosystem health [94].

PCBs, which are environmentally persistent and a “probable human carcinogen” as opposed to a “known carcinogen” [95], were above lower limits of quantitation in Fucus spp. at all our sites. While Σ₄₀PCBs concentrations at many sites exceeded the CSF-based screening level, they were still lower in an established portion size of seaweed than concentrations in other commonly consumed foods (Fig 4a, S3 Table). Even at Rock Bay, the site with the highest Σ₄₀PCBs, concentrations of PCBs per portion were only 17–29% of those reported in marine and freshwater fish, although they were higher than in other food types. One of the limitations to our approach is that we sum the 40 PCB congeners, where the CSF and RfD are based on total PCBs. To consider this limitation, we also used the West et al. [36] fish-based calculation for a subset of our congeners to estimate total PCBs (S1 Fig). The comparison to screening levels resulted in an additional four sites exceeding the RfD-based screening level and requiring a lowered recommended consumption rate (S4 Table). These sites and others would have exceeded the screening level had we used the LOQ instead of half the LOQ when concentrations were below the LOQ, demonstrating the uncertainty associated with the considerations of PCBs.

High molecular weight PAHs such as BaP or diseases caused by BaP exposure have also been monitored in English sole [94, 96] and two crustacean species: the Dungeness crab Metacarcinus magister and the spot prawn Pandalus platyceros [97]. These chemicals also have been reported in intertidal seagrasses (Zostera marina) near a shoreline oil refinery [98] and in developing embryonic Pacific herring transplanted near creosote-treated pilings in the nearshore, at levels high enough to impair their health [99].

Concentrations of BaP, classified as “carcinogenic to humans” [100], did not exceed the RfD-based screening level and only exceeded the USEPA CSF-based screening level (16 μg/kg) in two samples from Victoria Harbour: F. distichus at Rock Bay and F. spiralis at Point Hope (Fig 2b). When calculated as a concentration per portion, the high BaP concentrations in Rock Bay F. distichus made this sample a significant statistical outlier (Fig 4). As a result, BaP concentrations in F. distichus from Rock Bay were comparable to grilled, very well-done steaks and grilled, well-done chicken with bone and skin, foods with some of the highest reported BaP concentrations in commonly consumed foods (Fig 4b, S3 Table). Otherwise, average BaP concentrations/portion in F. distichus were lower than those in many commonly consumed foods (Fig 4b).

Concentrations of total As also frequently exceeded screening levels for inorganic arsenic, a “known human carcinogen” [95]. Furthermore, our diet analysis found that average total As concentrations in F. distichus and N. luetkeana exceeded concentrations per portion in commonly consumed foods except for marine fish (Fig 4c). Without knowing the relative proportions of organic and inorganic As species in our samples, the exceedances of the screening levels and their implications for consuming Salish Sea seaweeds are difficult to interpret. Studies that speciated As in seaweeds from BC found the inorganic As content to be 38% of total As in Fucus sp. [101] and 1–3% in N. luetkeana [102]. However, inorganic As in the genus Sargassum, including “Hijiki” which is sold commercially, are high enough that consumption advisories were issued regarding its use by the United Kingdom, New Zealand, Australia, China and Canada [103–105]. Sargassum muticum, an invasive, commonly harvested Salish Sea seaweed is also relatively high in inorganic arsenic [102]. Seaweed, fish, and shellfish are known to be high in arsenosugars [106]. Until recently, the toxicity of arsenosugars was thought to be of little concern [56, 101]. However, arsenosugars may be metabolized into potentially toxic metabolites, such as thio-dimethyl arsenic, that can have cytotoxic effects on the bladder and lungs [107, 108].

High levels of Cd can cause bone fragility, reduced lung function, and kidney damage [109]. Cd concentrations in Salish Sea F. distichus and N. luetkeana were uniformly below US RfD-based screening levels (Table 3), but consistently above international limits for seaweed (Table 1). Average Cd concentrations per portion in F. distichus and N. luetkeana were higher than in other commonly consumed foods (Fig 4d; S3 Table) but were lower than in shellfish from the Salish Sea. Concentrations of Cd in Salish Sea shellfish from BC range from 0.3 mg/kg wet weight (WW) in Manila clams (Venerupris phillipinarum) to 3.56 mg/kg WW in Miyagi oysters (Crassostrea gigas) [110]. When normalized by a typical portion size of 110 g per portion of uncooked shellfish [52], the resulting concentrations per portion of these shellfish (0.033–0.39 mg/portion) exceeded the highest concentrations per portion in F. distichus and N. luetkeana by up to tenfold (Fig 4d). High levels of Cd in cultured British Columbia oysters have led to the loss of some international trade markets [110, 111].

Pb exposure can alter neuromotor and neurosensory function, reduce cognitive function, hemoglobin levels, and kidney function, and cause developmental delays and reproductive impairment [112, 113]. Lead concentrations in our seaweed were only compared to international limits. In the US, the EPA has not assigned an RfD to Pb, but instead relies on the Integrated Exposure Uptake Biokinetic Model for Lead in Children (IEUBK). This Pb blood model is used to predict Pb concentrations in children, considered the most at-risk population. However, seaweed consumption data for children, which is needed to calculate these values, do not exist for our study. The concentrations in three Victoria/Esquimalt Harbour samples (F. distichus at Rock Bay and Point Hope, and F. spiralis at Portage Inlet) exceeded both the Chinese and French limits for Pb in seaweed (Table 3, Fig 2e). On a per portion basis, Pb in the two F. distichus statistical outliers in Victoria Harbour (Rock Bay and Point Hope) also far exceeded concentrations in commonly consumed foods (Fig 4e), but at all other sites they were comparable to concentrations in foods such as fruits, vegetables, and meats.

None of our samples exceeded the RfD-based screening level for methyl Hg (Table 2). Eighteen percent of F. distichus and 72% of N. luetkeana samples did exceed the French limit of 0.1 mg/kg (Fig 2f); however, the methodologies for choosing the French limit are not clear and the species of Hg in the French limit was not specified. Additionally, the relative proportions of different Hg species in our samples were not determined. Although most of the Hg in fish and seafood is methyl Hg [114], the proportion of methyl Hg in seaweeds is not well described [9]. Studies that measured total and methyl Hg in seaweed showed methyl Hg below limits of quantitation in Laminaria digitata and Saccharina latissima [115]; as 1.3 and 3.2 percent of total Hg in Ascophyllum nodosum and Fucus sp. [116]; and as 1.65% of total Hg in Fucus sp. [117]. Hg concentrations in F. distichus and N. luetkeana were comparable to those in many commonly consumed foods (Fig 4f). Even the seaweed sample with the highest total Hg concentration (N. luetkeana from Victoria Jetty) had concentrations per portion lower than those found in marine fish, such as tuna and salmon. Inorganic Hg exposure is of concern because it damages kidneys, stomachs, and intestines in children and adults, and can alter heart function in children when ingested in high amounts [118]. Methyl Hg causes neurologic damage, congenital disease, and immunosuppression, and affects the cardiovascular system [119].

The results of our study identify potential health risks from contaminants when consuming seaweeds harvested from the Salish Sea. The wide ranges of contaminant concentrations among sites suggests that there are certain sites where seaweeds should be avoided, eaten less frequently, or eaten in smaller portions to avoid exceeding CSF- or RfD-based limits, similar to the approach used to set consumption guidelines for fish [120]. As an example of this approach, we used our recalculated consumption rates for Salish Sea N. luetkeana, F. distichus and F. spiralis, which were based on concentrations of Cd, methyl Hg, Pb, BaP, and Σ₄₀PCBs measured in this study (Table 4).

This analysis shows that N. luetkeana can be consumed at all 18 sites at the typical portion size of 5 g DW/person/day for adults without exceeding our calculated screening levels based on USEPA RfDs for Cd or methyl Hg; USEPA CSFs for PCBs or BaP; or the Chinese limit for Pb. On the other hand, F. distichus can be consumed at an average rate of at least 5 g DW/person/day without exceeding these limits at only 34% of sites. These exceedances are driven primarily by PCBs; at 68% of sites, consuming F. distichus at 5/g DW/person/day would result in exceeding the CSF for Σ₄₀PCBs. Consuming 5 g DW/person/day of F. distichus from Point Hope would cause the CSF for Σ₄₀PCBs and the international limit for Pb to be exceeded and consuming it from Rock Bay in these amounts would cause CSFs for BaP and PCBs and the international limit for Pb to be exceeded. At sites for which only PCBs were problematic, recommended F. distichus consumption rates could be lowered; these modified consumption rates would range from 2.0 g DW/person/day at Vashon Island to 4.8 g DW/person/day at Hagan Bight. At sites where multiple contaminants exceeded regulated levels, the most conservative recalculated consumption rate could be used. For example, the recommended consumption rate of F. distichus would then be 1.9 g DW/person/day at Point Hope based on Pb concentrations and 1.2 g DW/person/day at Rock Bay based on BaP concentrations. Using a similar approach for F. spiralis, the recommended consumption rates would be 4.4 g DW/person/day at Fidalgo Bay based on the Σ₄₀PCB concentration and 1.9 g DW/person/day at Portage Inlet based on the Pb concentration. This approach does not account for seasonal variation in contaminant concentrations and would produce different recommendations if international limits were used. Furthermore, it does not apply to children.

For people who rely heavily on Salish Sea seafood in their diets, contaminants in seaweeds need to be considered within the context of total contaminant loads from all seafoods as well as seaweed health benefits and human dimensions (e.g. cultural well-being, sense of place, and availability of local foods for traditional, subsistence, and recreational use) that are not routinely considered in standard risk assessments. While Salish Sea Fucus and Nereocystis may absorb contaminants, they also contain many chemical constituents that may be beneficial to human health. Fucus and Nereocystis are high in fiber but low in fat, and contain essential minerals, such as potassium, iron, iodine, zinc, magnesium, and calcium ([73]; Table 3). They are also a source of antioxidants and phytochemicals, such as phlorotannins and fucoidans (Van Alstyne et al., unpublished data.), that have potential human health benefits [5]. Avoiding seaweed consumption because of concerns about contaminants could lead to making less healthy food choices by substituting foods that are low in fiber, high in sugar and unhealthy fats. Furthermore, the substituted foods may have similar contaminant levels (Fig 4). By choosing not to harvest or consume seaweed, people may also lose the benefit of strengthening cultural, family, and community connections and sharing traditional knowledge, in addition to the well-documented physical and psychological benefits of time spent outdoors in nature [121–125]. Thus, although this study focused on risks associated with contaminants in two species of Fucus and N. luetkeana in the Salish Sea, these risks should be considered along with the health benefits that result from consuming these seaweeds.