Subscriber access provided by CORNELL UNIVERSITY LIBRARY
Article
Temporal trends and geographical differences of perfluoroalkyl acids in Baltic Sea herring and white-tailed sea eagle eggs in Sweden Suzanne Faxneld, Urs Berger, Björn Helander, Sara Danielsson, Anna Aroha Miller, Elisabeth Nyberg, Jan-Olov Persson, and Anders Bignert Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.6b03230 • Publication Date (Web): 24 Oct 2016 Downloaded from http://pubs.acs.org on October 24, 2016
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Subscriber access provided by CORNELL UNIVERSITY LIBRARY
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 32
Environmental Science & Technology
(3, A)
(4, B, C)
(1)
(2)
ACS Paragon Plus Environment
Environmental Science & Technology
1
Temporal trends and geographical differences of
2
perfluoroalkyl acids in Baltic Sea herring and
3
white-tailed sea eagle eggs in Sweden
Page 2 of 32
4 5
Suzanne Faxneld1*, Urs Berger2†, Björn Helander1, Sara Danielsson1, Aroha Miller1,
6
Elisabeth Nyberg1, Jan-Olov Persson3, Anders Bignert1
7 8
1
9
History, Box 50007, SE-104 05, Stockholm, Sweden.
Department of Environmental Research and Monitoring, Swedish Museum of Natural
10
2
11
Stockholm, Sweden.
12
3
13
* corresponding author: e-mail:
[email protected], phone: + 46 (0) 8 5195 4114
Department of Applied Environmental Science, Stockholm University, SE-106 91,
Department of Mathematics, Stockholm University, SE-106 91, Stockholm, Sweden.
14 15 16
ABSTRACT
17
Temporal and spatial trends of perfluoroalkyl acids (PFAAs) were investigated in Baltic Sea
18
herring liver (Clupea harengus) from three sites, and white-tailed sea eagle (WTSE) eggs
19
(Haliaeetus albicilla) from two freshwater and two marine areas in Sweden. Trends of most
20
quantifiable PFAAs increased over the monitored period (1980 – 2014 in herring,
21
1960s/1980s – 2010 in WTSE). No significant decreasing trends were observed for the most
22
recent ten years for any substances, except perfluorooctane sulfonamide (FOSA).
23
Concentrations of perfluorooctane sulfonic acids (PFOS) in herring showed a distinct
1 ACS Paragon Plus Environment
Page 3 of 32
Environmental Science & Technology
24
decreasing trend moving from the more southern site towards the more northern site,
25
indicating main input of PFOS into the southern Baltic Sea. For WTSE, PFOS concentration
26
was higher in the marine compared to the freshwater environment, explained by the
27
cumulative historic contamination of the Baltic Sea. Similarly, concentrations in WTSE were
28
lower in the northern part of the Baltic Sea compared to further south. Concentrations of
29
PFUnDA, representing long-chain perfluoroalkyl carboxylic acids (PFCAs), showed a more
30
homogenous spatial distribution compared to PFOS for both herring and WTSE, indicating
31
that atmospheric inputs (via precursors) of the long-chain PFCAs are important contributors
32
in the study areas.
33 34
INTRODUCTION
35
Perfluoroalkyl acids (PFAAs) are a group of anthropogenic surfactants comprising a fully
36
fluorinated, hydrophobic alkyl chain typically consisting of 4 to 18 carbon atoms and a
37
hydrophilic functional group.1, 2 The two most well-known PFAAs are perfluorooctanoic acid
38
(PFOA) and perfluorooctane sulfonic acid (PFOS), both of which contain an eight carbon
39
atom backbone. PFAAs possess exceptional stability and surface tension-lowering ability1,
40
rendering them suitable for usage in many applications both industrially (e.g., production of
41
fluoropolymers) and in consumer products (e.g., water and stain proofing agents, fire-fighting
42
foams).1 Moreover, their stability makes them virtually non-degradable, hence
43
environmentally persistent.3 PFAAs have been produced since the early 1950s. In 2000, 3M, the main
44 45
company producing PFOS, voluntarily phased out PFOS, PFOA, and related chemicals.4
46
However, PFOS is still produced in Southeast Asia5, specifically in China.6 In 2009, PFOS
47
and related products were added to annex B (new persistent organic pollutants) of the
48
Stockholm Convention on Persistent Organic Pollutants, resulting in production and usage
49
restrictions.7 2 ACS Paragon Plus Environment
Environmental Science & Technology
Page 4 of 32
Today, PFAAs are ubiquitous environmental contaminants8 found in water9,
50 51
sediment10, biota11, 12, and humans.13 Long-chain PFAAs bioaccumulate in marine food
52
chains14, with highest concentrations found in organisms at higher trophic levels e.g., marine
53
mammals and predatory birds.11, 15, 8, 16 In contrast to lipophilic contaminants such as dioxins,
54
DDT, and PCBs, PFAAs bind to proteins17; hence highest concentrations are usually found in
55
protein-rich tissues such as liver, serum, and egg yolk.18 Herring (Clupea harengus) is a pelagic species that feeds mainly on
56 57
zooplankton. They are important for human consumption and as prey for marine predators,
58
including white-tailed sea eagles (WTSE) (Haliaeetus albicilla) that feed primarily on species
59
that prey on herring (e.g., pike, Esox Lucius, gulls, Laridae spp., and mergansers, Mergus
60
spp.). WTSE are large birds of prey. In the mid-1960s, they became critically
61 62
endangered in the Baltic due to PCB and DDT exposure.19, 20 After these compounds were
63
banned in the 1970s, populations slowly recovered. Currently, the WTSE population in
64
Sweden is approximately 700 pairs. While the majority of WTSE populations feed primarily
65
on fish and aquatic birds, some northern Lapland populations prey on large mammals such as
66
reindeer (Rangifer tarandus) carcasses during the ice-free period, and herbivores and seals
67
during winter. 45 In its position as apex predator, WTSE remain highly exposed to persistent
68
environmental contaminants. In Sweden, increasing concentrations of PFAAs have been
69
observed in common guillemot (Uria aalge) and peregrine falcon (Falco peregrinus) eggs21,
70
22
71
, grey seals (Halichoerus grypus)23, and otters (Lutra lutra).24 Here we examined temporal and spatial trends of PFAAs in Baltic Sea herring
72
and WTSE from Sweden. Trends are used to draw conclusions on emission histories and input
73
pathways of the different target analytes into the study areas.
74
3 ACS Paragon Plus Environment
Page 5 of 32
75
Environmental Science & Technology
MATERIALS AND METHODS
76 77
Sampling and sample preparation
78 79
Herring
80
Herring samples are collected annually from 17 sites along the Swedish coast within the
81
Swedish National Monitoring Programme for Contaminants in Marine Biota.25 Sampling
82
locations are regarded as reference sites, i.e. far from known pollution sources and ferry
83
routes. Individual herring are stored at -25oC in the Environmental Specimen Bank (ESB) at
84
the Swedish Museum of Natural History (SMNH). Fish sampling was conducted between
85
1980 and 2014 during the months of September to November. Collected specimens were
86
placed individually in polyethylene bags, frozen, and transported to the laboratory for
87
preparation.
88
For this study, three sites were chosen – Gulf of Bothnia (Ängskärsklubb),
89
northern Baltic Proper (Landsort), and southern Baltic Proper (Utlängan) (Figure 1) – because
90
herring from there was the oldest archived in the ESB. Samples stored in the ESB were used
91
for retrospective analysis of PFAA concentrations. Total body weight, liver weight, total
92
length, and gonad maturity were recorded. Herring age was determined via scales.51 Females
93
between 2 and 5 years of age and between 17 to 20 cm length were analysed. For the earlier
94
years (1980-2006), one pool of 12 individual liver samples was used for every other year and
95
site; from 2007 until 2014, two pools containing 12 individual liver samples per pool from
96
each site and year were analysed (see Table S5 for details about years analysed).
97 98
White-tailed sea eagles
4 ACS Paragon Plus Environment
Environmental Science & Technology
99
Page 6 of 32
Annual surveys of WTSE reproduction have been conducted in Sweden since the mid-
100
1960s26, involving collection of addled/dead eggs for examination. We included 83 single
101
eggs from different clutches from 1966 until 2010 from four regions (Figure 1); (1) 18 eggs
102
from 18 sites in northern Inland (Lapland); (2) 14 eggs from 13 southern Inland sites in
103
central Sweden; (3) 30 eggs from 15 sites in the Gulf of Bothnia; and (4) 21 eggs from 17
104
sites in the Baltic Proper. In total, 18 sites are represented by more than one clutch during the
105
study period with time spans from 1 to 40 years (mean 21 years) (see Table S6 for details
106
about years analysed). Each egg was stored in a polyethylene bag inside a box/beaker for
107
transport to the laboratory at the SMNH. Egg contents were homogenized and stored
108
individually at -80o C in the ESB at the SMNH. Individual samples were taken for analysis of
109
PFAAs.
110
WTSE eggs were collected 1 to 2 months after failure to hatch, and were in
111
varying states of desiccation. Corrected wet weight data for PFAAs were calculated by
112
multiplying the measured concentration in the homogenized sample with a desiccation index
113
value based on the sampling weight of the individual egg, related to its interior volume20, and
114
by multiplying with a factor of 1.1 to roughly compensate for the air sac within the egg.
115 116
5 ACS Paragon Plus Environment
Page 7 of 32
Environmental Science & Technology
f fo ul G
A
Ba
B
lti c
Pr
op er
3
ia hn t Bo
1 Inland north Lapland (Lap)
2 Inland central and southern regions (Inl) 3
South Bothnian Sea (sBS)
4
Baltic Proper (BP)
C 117 118
Figure 1. Map of sampling areas. 1-4 are white-tailed sea eagle sites. 1=northern Inland (Lapland), 2=southern
119
Inland, 3=Gulf of Bothnia, 4= Baltic Proper. A-C represent herring sites; A= Ängskärsklubb (Gulf of Bothnia),
120
B= Landsort, C= Utlängan (both in the Baltic Proper).
121 122
Analytical methods
123
Abbreviations of PFAAs (see supplementary information) are according to Buck et al.2 The
124
target analytes in this study were PFBS, PFHxS, PFOS, PFDS, FOSA, PFHxA, PFHpA,
125
PFOA, PFNA, PFDA, PFUnDA, PFDoDA, PFTrDA, PFTeDA, and PFPeDA. FOSA was the
126
only precursor compound included here and will, for simplicity, be included under the generic
127
term PFAAs. The sample extraction procedure used has been described previously.27 Briefly,
128
a sample aliquot of approximately 1.0 g (herring liver) or 0.5 g (WTSE egg) homogenized
129
tissue in a polypropylene (PP) centrifuge tube was spiked with 1.0 ng (herring) or 10 ng 6 ACS Paragon Plus Environment
Environmental Science & Technology
Page 8 of 32
130
(WTSE) each of a suite of mass-labelled internal standards (18O- or 13C-labelled
131
perfluoroalkane sulfonic acids [PFSAs] and perfluoroalkylcarboxylic acids [PFCAs]).
132
Samples were extracted twice with 5 mL of acetonitrile in an ultrasonic bath. Following
133
centrifugation, the supernatant extract was removed and the combined acetonitrile phases
134
were concentrated to 1 mL under a stream of nitrogen. The concentrated extract underwent
135
dispersive clean-up on graphitised carbon and acetic acid. A volume of 0.5 mL of the purified
136
extract was added to 0.5 mL of aqueous ammonium acetate. Precipitation occurred and the
137
extract was centrifuged before the clear supernatant was transferred to an autoinjector vial for
138
instrumental analysis, and volume standards 13C8-PFOA and 13C8-PFOS were added.
139
Aliquots of the final extracts were injected automatically on an ultra-performance liquid
140
chromatography (UPLC) system (Acquity, Waters) coupled to a tandem mass spectrometer
141
(MS-MS; Xevo TQS, Waters). Instrumental analysis and quantification was performed
142
according to Vestergren et al.28 In short, compound separation was achieved on a BEH C18
143
UPLC column (1.7 µm particles, 50× 2.1 mm, Waters) with a binary gradient of ammonium
144
acetate buffered methanol and water. The MS-MS was operated in negative electrospray
145
ionisation mode. Quantification was performed in selected reaction monitoring
146
chromatograms using the internal standard method. All results are given on a sample wet
147
weight (ww) basis. Quantified concentrations for PFAAs in WTSE eggs were corrected for
148
the respective egg’s desiccation (see above) before statistical analysis.
149 150
Quality control
151
The extraction method used here (with the exception of the concentration step) has previously
152
been validated for biological matrices and showed excellent analyte recoveries ranging
153
between 90 and 110% for PFCAs from C6 to C14.29 Including extract concentration, we
154
determined recoveries between 70 − 90% for C6- to C10-PFCAs and 65 − 70% for C11-C15
7 ACS Paragon Plus Environment
Page 9 of 32
Environmental Science & Technology
155
PFCAs. Extraction efficiencies for PFSAs, including FOSA, were determined to be 70 − 95%.
156
Method quantification limits (MQLs) for all analytes were determined on the basis of blank
157
extraction experiments and ranged between 0.01 and 0.3 ng/g ww. Fish tissue sample used in
158
an international inter-laboratory comparison (ILC) study in 200730 was analysed as a control
159
along with all sample batches. The concentrations obtained were in good agreement with
160
mean concentrations from the ILC study for all seven compounds quantified in the ILC.
161 162
Statistical analysis
163
Values below Method Quantification Limits (MQL) were replaced by MQL divided by the
164
square root of 2 prior to all statistical analyses. Power analyses were carried out to estimate
165
the lowest detectable trend for a monitoring period of 10 years at a fixed power of 80%, given
166
the current residual variance from a log-linear regression analysis. In the same way, the
167
number of years required to detect an annual trend of 10% were estimated for the various data
168
sets. The significance level (α) for all statistical tests was 5%. Data were log transformed
169
before analysis, and yearly geometric means were calculated (Table S5, S6). We used
170
statistical software PIA52 for trend analyses.
171
ΣPFSA was calculated and included the detected PFSAs - PFHxS and PFOS for
172
herring and PFHxS, PFOS, and PFDS for WTSE. ΣPFCA was calculated. For both species it
173
included PFOA, PFNA, PFDA, PFUnDA, PFDoDA, and PFTrDA, and for WTSE also
174
PFTeDA and PFPeDA. To study changes in homologue pattern between different sampling
175
periods and species, Principal Component Analysis (PCA) was performed on the proportions
176
of a set of individual PFAAs relative to the sum of this set, and log-transformed prior to PCA
177
analysis. Before the PCA-scores were plotted they were centered and scaled to 100%. The
178
PCA loadings were overlaid as a biplot and Hotellings 95% confidence ellipses for the center
179
of gravity for each site plotted. We conducted a Hotellings T-square test to detect possible
8 ACS Paragon Plus Environment
Environmental Science & Technology
180
significant differences in homologue pattern (multivariate means) between the different
181
sampling periods. Because three to four time-periods were tested against each other, p-values
182
were compared using a Bonferroni adjusted significance level of 0.0167. Tests were carried
183
out using PIA.52
184
Page 10 of 32
We performed a log-linear regression for the temporal trend analysis for the
185
entire investigated period and the most recent 10 years using yearly geometric mean values.
186
To achieve a fairly stable estimate of the recent concentrations, a geometric mean of the last 3
187
years was calculated. A simple 3-point running mean smoother was fitted to the annual
188
geometric mean values. Additionally, non-linearity of trends was investigated using a
189
Change-Point detection method suggested by Sturludottir et al.31 This method iteratively
190
searches for a combination of two log-linear regression lines that explains significantly more
191
of the total variance than that explained by a single regression line for the whole study period.
192
In order to combine temporal trends at various contaminant levels in the same graphs, data
193
were transformed to percent of maximum concentration prior to change-point detection
194
analysis, whereas the tables (S1, S2) contains the log-transformed values, expressed as
195
geometric mean.
196
In order to investigate spatial trends of PFAA in herring and WTSE, Stata13
197
was used. PFOS and PFUnDA were investigated as representative of long-chain PFSAs and
198
PFCAs, respectively, because they showed the highest detection frequencies in the analysed
199
samples. For PFOS, we investigated the period 1980 to 2000 because phase-out of PFSAs by
200
3M began in 2000, which significantly affected subsequent trends. For PFUnDA, 1980 to
201
2010 was used, because samples from this period were available for almost all sampling sites
202
for both species. If the slopes of the time trends of PFAA concentrations between the
203
sampling sites for a given species did not differ significantly, spatial differences in PFAA
204
concentrations were investigated. This was done using estimated concentration values from
9 ACS Paragon Plus Environment
Page 11 of 32
Environmental Science & Technology
205
the time trend regression for PFOS and for PFUnDA in 1990 and 1995, respectively (middle
206
of the respective period). For WTSE spatial analyses, the Gulf of Bothnia and the Baltic
207
Proper were grouped as marine areas, while the southern and northern inland areas were
208
grouped as freshwater areas.
209 210 211
RESULTS AND DISCUSSION
212
Henceforth, PFOS and PFUnDA are used as examples of PFSA and PFCA, respectively,
213
because PFOS had the highest concentrations and has now been phased out, and because
214
PFUnDA concentrations were uniformly detected in both herring and WTSE. Detailed results
215
for all compounds are included in the supplementary data.
216
For herring, 9 of 15 analysed PFAA were detected in most samples from all
217
three sites. PFBS, PFDS, PFHxA, PFHpA, PFTeDA, and PFPeDA were below MQL in >
218
50% of investigated years, and were thus not included in the statistical analysis. In WTSE
219
eggs, 12 of 15 investigated PFAA were detected; however, PFOA was infrequently detected
220
in samples from the two inland areas. PFBS, PFHxA, and PFHpA were generally below MQL
221
and were thus excluded from statistical analysis.
222 223
Concentrations and contaminant profiles
224
PFOS is typically the dominant PFSA in wildlife.12 Here, PFOS was the dominant compound
225
in both herring and WTSE at all sites and in all years (Figure S1-S7). In herring, PFOS
226
percentage increased over time, constituting 80-90% of ΣPFSA in the 1980s to 1990s, to
227
around 95% post-1990s. In WTSE, PFOS was more stable throughout the examined time
228
period, constituting approximately 98% of ΣPFSA.
229 230
In herring, PFOS concentrations in the most recent year i.e., 2014, were similar at the two Baltic Proper sites i.e., Landsort and Utlängan (13-14 ng/g ww, sites B and C, 10 ACS Paragon Plus Environment
Environmental Science & Technology
Page 12 of 32
231
Figure 1), while concentrations were half that at Ängskärsklubb, in the Gulf of Bothnia (site
232
A, Figure 1) (Table S1). In WTSE, concentrations of PFOS in 2010 were similar in the Gulf
233
of Bothnia and the southern inland samples, while concentrations in the Baltic Proper were
234
twice as high. By contrast, northern Inland had concentrations three times lower compared to
235
the Gulf of Bothnia and the southern Inland site (Table S2).
236
For ƩPFCA, PFOA, PFNA, and PFUnDA were dominant in herring,
237
contributing 20-22%, 28-30%, and 19-22%, respectively (Figure S1-S3), consistent with
238
results from slimy sculpin (Cottus cognatus) in Lake Ontario, where PFOA and PFUnDA
239
were also the dominant PFCAs.32 In skipjack tuna (Katsuwonus pelamis) from the mid-north
240
Pacific and Indian Oceans, PFUnDA was the dominant PFCA, while PFOA was detected in
241
only a few samples.46 In WTSE, the dominant PFCAs were PFNA, PFUnDA, and PFTrDA,
242
constituting 20-23%, 24-28% and 18-30% of ƩPFCA, respectively (Figure S4-S7). Likewise,
243
PFUnDA and PFTrDA were the dominant PFCAs in peregrine falcon eggs22, Baltic common
244
guillemot eggs33 in Sweden, and in ancient murrelet (Synthliboramphus antiquus), Leach’s
245
storm-petrel (Oceanodroma herodias), and rhinoceros auklets (Cerorhinca monocerata) from
246
the western Pacific coast of Canada.34 Moreover, in double-crested cormorants
247
(Phalacrocorax auritus) and great blue herons (Ardea herodias), two coastal bird species,
248
PFNA was the dominant PFCA.34 The relative amount of PFOA was much lower in WTSE
249
compared to herring, constituting only about 1.5-3%. Such large differences in PFOA
250
contribution to ƩPFCA between predatory bird and prey fish has been described earlier,47 and
251
could be due to differences in bioaccumulation factors based on differences in toxicokinetics,
252
especially elimination half-lives, or due to precursor degradation along the food chain.47
253
PFOA percentage of ƩPFCA decreased over time in WTSE from all areas over
254
the entire time period (i.e. 1960s/1980s to 2010) and also in herring from the Gulf of Bothnia
255
(1980 to 2014). For PFUnDA, increases over the entire time period were seen in herring from
11 ACS Paragon Plus Environment
Page 13 of 32
Environmental Science & Technology
256
the Gulf of Bothnia and northern Baltic Proper, and in WTSE from the Baltic Proper and
257
northern Inland. In herring, PFUnDA constituted 12-15% in the 1980s, increasing to 20-25%
258
in the 2000s. In WTSE from the Baltic Proper, PFUnDA constituted 5-15% in the 1970s,
259
increasing to approximately 20% in the 2000s. PFTeDA, PFPeDA, and PFDS were found
260
only in WTSE samples (Tables S5, S6). In melon-headed whales (Peponocephala electra)
261
from Japan, the concentration of PFUnDA increased greatly from 1982 to 2006.48 In contrast,
262
in northern sea otters (Enhydra lutris kenyoni) from Alaska, PFUnDA was rarely detected and
263
only found in low concentrations, while PFNA was the dominant PFCA, increasing from
264
1992 to 2007.49
265
In WTSE, concentrations of odd chain PFCAs (PFNA [C9], PFUnDA [C11], and
266
PFTrDA [C13]) were much higher compared to their adjacent even chain PFCAs (Table S2), a
267
pattern not seen in herring. Similarly, this pattern of higher concentrations of long-chained
268
odd numbered PFCAs has been reported in various species e.g., rhinoceros auklet, Leach’s
269
storm petrels, and ancient murrelet34, beluga whales (Delphinapterus leucas) from Alaska,
270
herring gull eggs from Norway, and seabird eggs from the Canadian Arctic.35, 36, 37 Moreover,
271
in otters this pattern was more pronounced for marine compared to freshwater samples.24
272
Such a pattern in biota indicates a significant contribution of PFCAs originating from
273
atmospheric transformation of precursor compounds.38 However, this process alone could not
274
explain the difference in patterns observed between the freshwater and marine WTSE samples
275
in this current study. The marine and freshwater environments of Sweden represent different
276
emission histories (cumulative historic or snapshot in time, respectively22), which will also
277
have an influence on the detected pattern, as discussed below.
278
A clear change in the relative concentration of the PFAAs over time in herring
279
was observed (Figure 2a). PFHxS and FOSA dominated in the earlier years (1983-1999),
280
while between 2000-2009 and 2010-2014 the relative concentrations of some of the
12 ACS Paragon Plus Environment
Environmental Science & Technology
Page 14 of 32
281
carboxylic acids, PFNA, PFUnDA, and PFTrDA, increased. There were also some significant
282
differences between the three herring sites. For 1983-1999, the relative concentration of
283
FOSA was higher in the Gulf of Bothnia, while in the southern Baltic Proper, the relative
284
concentration of PFOS was higher. In 2010-2014, the relative concentration of PFUnDA was
285
higher in the Gulf of Bothnia compared to the southern and northern Baltic Proper. No
286
significant differences between sites were seen in 2000-2009. For WTSE, there were
287
significant differences in relative concentration of PFAAs between the different time periods,
288
except 1980-1989 and 1990-1999 (Figure 2b). Earlier years were dominated by PFOS and
289
PFHxS while in the later years, relative concentrations of the carboxylic acids PFNA,
290
PFUnDA, and PFTrDA, showed increased relative concentrations. In addition, for the later
291
time period, i.e. 2000-2010, the relative concentrations of PFHxS and FOSA were lower in
292
the southern Inland area compared to the other three WTSE areas. A comparison of the
293
PFAAs pattern in herring and the Baltic Sea WTSE (i.e. Gulf of Bothnia and Baltic Proper
294
areas) in 1980-2010 showed significant differences (Figure 2c). Relative concentrations of
295
PFOS and PFTrDA were higher in WTSE, whereas the relative concentrations of PFHxS,
296
FOSA, and PFNA were higher in herring.
297 298
a)
13 ACS Paragon Plus Environment
Page 15 of 32
Environmental Science & Technology
299 300 301
b)
302 303 304 305
c)
14 ACS Paragon Plus Environment
Environmental Science & Technology
Page 16 of 32
306 307
Figure 2. Principal component analysis, biplot, and Hotelling’s 95 % confidence ellipses for centers of gravity
308
for each group. Changes over time in relative abundance of PFOS, PFHxS, FOSA, PFNA, PFUnDA, and
309
PFTrDA in a) herring at three different sites, Gulf of Bothnia (GB), northern Baltic Proper (nBP), and southern
310
Baltic Proper (sBP) and divided into three time periods, 1983-1999, 2000-2009, and 2010-2014 b) WTSE at four
311
different areas, Gulf of Bothnia (GB), Baltic Proper (BP), southern Inland (SI), and northern Inland (NI) and
312
divided into four time periods, 1966-1979, 1980-1989, 1990-1999, and 2000-2010 c) differences in PFAAs
313
pattern between herring and Baltic Sea WTSE in 1980-2010.
314 315
Temporal trends of PFSAs
316
Generally, increasing trends of PFSA compounds over the whole time period were seen for
317
both herring and WTSE, but not at all sites/areas. FOSA showed decreasing concentrations at
318
some herring sites (Table 1, Figure 3, Tables S1-S2, Figures S8, S9).
319 320
Table 1. Concentration trends (average annual percentage change) for PFOS and PFUnDA (herring liver and
321
WTSE eggs) assessed from annual geometric mean values.
Herring
Compound
Site name
Ntot
PFOS
s. Baltic Proper
31
Nyrs
Year
Trend% 95% CI
P
23
1980-2014
4.4 (2.8,6.0)
0.000 +++
10
2005-2014
3.2 (-3.3,10)
0.2991
15 ACS Paragon Plus Environment
Page 17 of 32
Environmental Science & Technology
PFUnDA
WTSE
PFOS
n. Baltic Proper
31
Gulf of Bothnia
31
s. Baltic Proper
31
n. Baltic Proper
31
Gulf of Bothnia
31
Baltic Proper
21
Gulf of Bothnia
southern Inland
northern Inland
PFUnDA
30
14
18
Baltic Proper
21
Gulf of Bothnia
30
southern Inland
14
northern Inland
18
23
1980-2014
7.3 (5.3,9.2)
0.000 +++
10
2005-2014
3.9 (-3.6,12)
0.2752
23
1980-2014
5.9 (3.9,7.9)
0.000 +++
10
2005-2014
-.16 (-10,11)
0.9239
23
1980-2014
7.3 (5.4,9.1)
0.000 +++
10
2005-2014
-5.2 (-14,4.0)
0.2206
23
1980-2014
8.6 (6.9,10)
0.000 +++
10
2005-2014
-2.9 (-8.7,3.2)
0.3009
23
1980-2014
8.5 (6.4,11)
0.000 +++
10
2005-2014
-4.6 (-16,8.0)
0.4125
15
1966-2010
6.9 (4.2,9.7)
0.000 +++
6
2001-2010
-2.5 (-12,7.6)
0.520
24
1969-2010
7.2 (4.5,9.9)
0.000 +++
8
2001-2010
-7 (-28,20)
0.513
11
1986-2010
1.5 (-5.5,9.1)
0.644
5
2002-2010
-3.6 (-30,32)
0.727
14
1966-2009
.89 (-2.8,4.7)
0.618
6
2001-2009
-8.3 (-39,38)
0.587
15
1966-2010
15 (12,19)
0.000 +++
6
2001-2010
4.8 (-10,22)
0.448
24
1969-2010
13 (10,16)
0.000 +++
8
2001-2010
-7.1(-24,13)
0.396
11
1986-2010
6.3(.98,12)
0.024 +
5
2002-2010
-3.9(-27,27)
0.678
14
1966-2009
12(9.6,14)
0.000 +++
6
2001-2009
8.4(-3.4,22)
0.122
322
The total number of samples for the whole time period (Ntot) and number of years (Nyrs) for the various time-
323
series are given. A second line at each site presents the results for the last 10 years period. P shows the p-value of
324
the trend. A + after the p-value indicates a positive trend; + p
