Chemical analyses of dredged spoil disposal sites at the Belgian part of the North Sea

Chemosphere 156 (2016) 172e180

Contents lists available at ScienceDirect

Chemosphere journal homepage: www.elsevier.com/locate/chemosphere

Chemical analyses of dredged spoil disposal sites at the Belgian part of the North Sea Bavo De Witte a, *, Ann Ruttens b, Bart Ampe a, Nadia Waegeneers b, Johanna Gauquie a, Lisa Devriese a, Kris Cooreman a, Koen Parmentier c a b c

Institute of Agricultural and Fisheries Research, Animal Sciences Unit e Aquatic Environment and Quality, Ankerstraat 1, 8400 Ostend, Belgium Veterinary and Agrochemical Research Center (CODA-CERVA), Trace Element Service, Leuvensesteenweg 17, 3080 Tervuren, Belgium Royal Belgian Institute of Natural Sciences, OD Nature, Ecochem, 3de en 23ste Linieregimentsplein, 8400 Oostende, Belgium

h i g h l i g h t s
PCB concentrations in Belgian marine sediments not decreasing since 2005.
Zn and Cu concentrations at dredged spoil disposal sites possibly affected by shipyards.
Dredged spoil disposal site monitoring essential to identify adverse time trends.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 27 December 2015 Received in revised form 5 April 2016 Accepted 30 April 2016 Available online 10 May 2016

The chemical status of five dredged spoil disposal sites in the Belgian Part of the North Sea is evaluated. A linear mixed-effect model was applied to PCB, PAH and heavy metal data from 2005 to 2014. No decrease in PCB concentrations was found, with even an increase at two disposal sites. Hg/AL ratios increased with 62% at one disposal site (BR&WS2) from 2005 to 2006 to 2013e2014. Cu and Zn concentrations increased at two disposal sites. Additional harbour sampling suggests that the latter is possibly linked to antifouling paints. Based on OSPAR environmental assessment criteria, the current chemical status of the sites suggests no chronic effect of dredged spoil disposal. However, increasing time trend data for PCB, Hg, Cu and Zn demonstrate the importance of monitoring to identify adverse trends. © 2016 Elsevier Ltd. All rights reserved.

Handling Editor: Caroline Gaus Keywords: Dredged spoil disposal Heavy metals PCB PAH Pollution control measures

1. Introduction Marine dredging can be divided in maintenance dredging works (to safeguard navigation in ports, harbours and channels), and capital dredging works (to deepen channels or ports, construct and burry cables). When strict quality criteria are met, dredged sediment can be disposed at sea. As dredged spoil disposal is potentially harmful to the marine environment, a strict monitoring of

* Corresponding author. E-mail addresses: [email protected] (B. De Witte), Ann.ruttens@ coda-cerva.be (A. Ruttens), [email protected] (B. Ampe), Nadia. [email protected] (N. Waegeneers), [email protected] (J. Gauquie), [email protected] (L. Devriese), Kris.cooreman@ilvo. vlaanderen.be (K. Cooreman), [email protected] (K. Parmentier). http://dx.doi.org/10.1016/j.chemosphere.2016.04.124 0045-6535/© 2016 Elsevier Ltd. All rights reserved.

biological and chemical impact is obliged (OSPAR, 2009a). On the Belgian Part of the North Sea (BPNS), five dredged spoil disposal sites are defined, and monitored two times a year. The current chemical monitoring of dredged spoil disposal sites at the BPNS routinely includes measurements of polychlorobiphenyls (PCB), polycyclic aromatic hydrocarbons (PAH) and heavy metals (Al, Fe, Cd, Cr, Cu, Hg, Ni, Pb and Zn). Monitoring of these compounds is also included in the European Marine Strategy Framework Directive (MSFD 2008/56/EC). Within descriptor 8, it is described that these pollutants may not give rise to pollution effects. Within MSFD, chemical concentrations in sediments of the BPNS are assessed against OSPAR environmental assessment criteria (EAC). Large amounts of PCB were manufactured between 1930 and 1983 and reach the environment through disposal, leakage,

B. De Witte et al. / Chemosphere 156 (2016) 172e180

evaporation and accidents (Roose et al., 2005). Although PCB were banned in the mid-1980s, sources still remain, e.g. within waste disposal, PCB-containing equipment, by remobilisation from sediments or by formation of by-products in thermal and chemical processes (OSPAR, 2012). Contradictory information can be found on PCB time trends at the BPNS after 1990. Roose et al. (2005) did not find a significant decrease from 1991 to 2001 for the PCB CB153 (no decrease of more than 20%). Due to good concentration correlations between various marker PCB, the authors considered CB153 representative for other marker CBs. The OSPAR quality status report 2010 (OSPAR, 2010a) indicates no significant decrease for more than 80% of PCB time series from 1998 to 2007 in OSPAR zone II, including the BPNS. Assessment was done by identifying linear trends in the log concentration over a 10 years period at the 5% significance level. In contrast, Everaert et al. (2014) report a decrease of 50e66% in the BPNS between 1991 and 2010 while concentrations at the nearby Scheldt estuary revealed no decreasing trend. PAH may enter the marine environment by offshore activities, oil spills, river discharges and atmospheric transport (OSPAR, 2012). A predominantly downward trend is noticed in biota throughout the OSPAR area, but is less pronounced in marine sediments (OSPAR, 2010a). Heim and Schwarzbauer (2013) estimated an increase of PAH fluxes since 1996 due to an increased energy consumption. PAH levels in the Scheldt estuary did not decrease in 2000e2009 (Gao et al., 2013). Stringent pollution control measures have led to substantial reductions in heavy metal pollution by industrial combustion processes, metal production, transport and waste streams (OSPAR, 2010a). OSPAR reports a strong reduction in biota and sediments during the 1990s followed by a slowdown (OSPAR, 2010a). This is consistent with data from the Scheldt estuary and Belgian coastal zone reported by Gao et al. (2013) and Baeyens et al. (2005), covering the time window of respectively 1978e1998 and 1985e2000. In the present study, time series of PCB, PAH and heavy metals, obtained from dredged spoil disposal site monitoring in the BPNS, are analysed by a linear mixed-effect model. In addition to the identification of significant time trends, the aim is to evaluate if trends are affected by dredged spoil disposal. Additional harbour monitoring was performed to estimate the effect of antifouling paints from leisure boats on Cu and Zn contamination. Since the ban of tributyltin (TBT), Cu made a revival as an antifouling, while several new products (Irgarol, Zn Pyrithione) contain Zn. 2. Materials and methodology 2.1. Description of sampling sites The BPNS contains five disposal sites (Fig. 1). Near the port of Zeebrugge, disposal sites BR&WS1, BR&WS2 and BR&WZE equal an area of 7.07, 0.88 and 1.77 km2 with an average dredged spoil disposal rate of 0.78, 2.05 and 1.79 ton dry matter (ton DM).m2.y1 from April 2004 to March 2014, respectively. BR&WOO near port Oostende and LNP near port Nieuwpoort receive less dredged spoil, respectively 0.36 and 0.09 ton DM.m2.y1 for an area of 1.77 km2 each (Lauwaert et al., 2014). From April 2004 to March 2014, all disposal sites mainly received maintenance dredged spoil, ranging from 83% (BR&WS1) to 100% (LNP). With exception of Br&WS2, there are no large fluctuations in yearly disposal rates (Table S1). Differences in chemical load between disposal sites can be expected since BR&WS1, BR&WS2 and BR&WZE mainly receive dredged spoil from the industrialized port of Zeebrugge (42.5 million tons of freights in 2014, Anon, 2015a) and the shipping track to the Western Schelde

173

which connects the port of Antwerp (199 million tons of freights in 2014, Anon, 2015b). At port Oostende, 1.4 million tons of freights were handled in 2014 (Anon, 2015c) while this port also contains a significant leisure boat harbour. Nieuwpoort is especially a leisure boat harbour. 2.2. Sampling and sample preservation Sediment samples at the BPNS were taken two times a year, i.e. in February/March and September/October. In this publication, results of PCB and heavy metal data from 2005 to 2014 are included. PAH time series are limited since systematic data are only available since 2008. Five disposal sites (BR&WS1, BR&WS2, BR&WZE, BR&WOO, LNP) are monitored with three zones for each disposal site: (1) the actual disposal site (DIS), (2) the directly impacted zone (IMZ), i.e. outside but less than 0.3 nautical mile away from the actual disposal site and (3) reference samples taken on longer distance from the disposal site (REF). Trend modelling (x2.4) includes the three zones and was performed on the raw data. In the discussion, focus is on the actual disposal site: average concentrations, increase or decrease percentages are given for DIS, unless otherwise mentioned. For heavy metals, changes in concentration were discussed, relative to the Al concentration. A map locating disposal sites and sampling points is shown in Fig. 1. Coordinates of sampling locations are provided as supplementary information. All samples were taken by a Van Veen grab (0.1 m2) and immediately deep frozen (20  C) on board the research vessel Belgica. In the lab, samples were wet sieved on a 63 mm sieve, followed by freeze drying and ball milling. The proportion of the <63 mm fraction is on average 5.1, 1.3, 45.2, 5.0 and 7.1% for BR&WS1, BR&WS2, BR&WZE, BR&WOO and LNP, respectively. Additional sampling was performed at the harbours of Nieuwpoort (HNP) and Oostende (HOO) in November 2013 and April 2014. Samples were taken by small Van Veen grabs (0.03 m2) from the rigid inflatable boat Zeekat (Nieuwpoort) or from the quaysides (Oostende). Fig. 1 gives an overview of the sampling locations, while coordinates are provided in Table S3. 2.3. Contaminant analysis For the analysis of PCB, the sum of 7 PCB, suggested by OSPAR for environmental monitoring (OSPAR, 2010b), was made. This sum includes IUPAC numbers CB28, CB52, CB101, CB118, CB138, CB153 and CB180. Two distinct methods were applied. All samples until 2012 were analysed by a Soxhlet extraction on freeze dried sediment with hexane:acetone (3:1) for 6 h at 90  C. This was followed by desulfurization by tetrabutylammonium sulphite, dissolved in deionized water. After phase separation, the solvent phase was dried by Na2SO4. Clean-up was done on a column with deactivated aluminium oxide as stationary phase and hexane as mobile phase, followed by fractionation on a glass column with silicon oxide and hexane as mobile phase. Tetrachloronaphthalene was added as an internal standard before splitless injection at 210  C on a gas chromatograph with electron capture detection (GC-ECD). Separation was done on a capillary HT8 column (SGE, 50 m, 0.22 mm, 0.25 mm) with hydrogen (Parker hydrogen generator) as carrier gas and with a temperature program starting at 90  C for 2 min, up to 180  C at 20  C min1 to 285  C at 2.8  C min1 which was held for 10 min. Flow rate was constant at 0.8 ml min1. Detector was set on 300  C with 25 ml min1 Ar-CH4 (95:5) as make-up gas. Freeze dried samples from 2013 and later were analysed by pressurized liquid extraction (PLE) at 100  C with hexane:acetone (3:1) as extraction solvent. Florisil was added to the cell for clean up. CB29, CB112 and CB140 were added in the extraction cells as calibration standards. Three extraction cycles of 5 min static time

174

B. De Witte et al. / Chemosphere 156 (2016) 172e180

Fig. 1. Overview of sampling locations at (a) harbour Nieuwpoort, (b) harbour Oostende and (c) dredged spoil disposal sites at the BNPS.

B. De Witte et al. / Chemosphere 156 (2016) 172e180

each were programmed. The extract was desulfurized by the addition of activated copper and fractionized on a solid phase extraction silicon oxide column. Tetrachloronaphtalene was added as an internal standard before splitless injection on a GC-ECD, analogous to the previous PCB method except for the oven program: starting at 90  C for 2 min, the temperature rises at 20  C min1 to 140  C, at 1  C min1 to 200  C and at 2.8  C min1 to 285  C which was held for 10 min. For PAH analysis, the sum of the 16 priority PAH, as defined by the United States Environmental Protection Agency (US-EPA), was compared (Donata, 2010). These include naphthalene, acenaphtylene, acenaphtene, fluorene, phenanthrene, anthracene, fluoranthene, pyrene, benzo(a)anthracene, chrysene, benzo(b) fluoranthene, benzo(k)fluoranthene, benzo(a)pyrene, indeno(1,2,3c,d)pyrene, dibenzo(a,h)anthracene and benzo(g,h,i)perylene. Freeze dried sediment was extracted with hexane:acetone (3:1) at 100  C by PLE. Florisil was added to the cell as clean up. Deuterated standards acenapthene d10, anthracene d10, pyrene d10, benzo(a)anthracene d12, benzo(a)pyrene d12 and indeno(123cd)pyrene d12 were added to the extraction cells. Three extraction cycles of 5 min static time each were programmed. Samples were analysed by a GC-mass spectrometer (GC-MS), equipped with a PTV-injector with baffled liner, temperature programmed from 80  C to 320  C. Separation was done on a select PAH column (Agilent, 30 m, 0.25 mm, 0.25 mm) with helium (Air Liquide, Alphagaz 2) as a carrier gas at a constant flow (2.2 ml/min). The oven temperature was programmed at 70  C for 0.79 min, followed by an increase of 35.5  C min1 to 180  C, 3.55  C min1 to 230  C, held for 13.8 min, 25.36  C min1 to 280  C, held for 13.8 min and a final 15.22  C min1 increase to 325  C which was held for 2 min. Detection was done by electron impact ionisation in single ion mode with transfer line temperature at 340  C, ion source temperature at 300  C and quadrupole temperature at 150  C. Freeze-dried sediment samples for heavy metal analysis were digested in teflon bombs with a mixture of concentrated HClO4, HNO3 and HF at 170  C for 16 h. Dry residues were redissolved in HNO3 prior to analysis with ICP-OES (Zn, Pb, Ni, Cr, Cu, Fe, Al) or ICPMS (Cd). For the determination of Hg, a dry combustion with oxygen was performed, followed by Au-adsorption and CV-AASquantitation. For the analysis of total organic carbon (TOC), sediment samples are mixed with a dichromate-sulphuric acid mixture, and titrated with 0.2 N Mohr’s salt with N-phenylanthranilic acid as indicator. All sequences included positive control analyses, solvent and procedure blanks. Proficiency testing was done for all analysis by yearly participation at the Quasimeme interlaboratory exercises (Quasimeme, Wageningen, The Netherlands), focused on marine samples. The analyses of PAH are ISO/IEC 17025 accredited. Quantification limits are given as Supplementary Information. 2.4. Statistical analysis Statistical analysis was done by a linear mixed-effect model in R (Version 2.15.3). The model included the factors time, time2, dredged spoil disposal site (SDS), season and the interaction time:SDS. Normality was evaluated for each model by a histogram and qq-plot of the residuals. Time and SDS were always included in the model as the objective was twofold: (1) the identification of significant time trends and (2) the effect evaluation of dredged spoil disposal. All other factors were removed one by one when not significant (a ¼ 0.05), starting with the interaction time:SDS. The factor season was included since seasonal effects may occur between samples taken in spring or autumn. The factor SDS distinguishes three zones previously defined: (1) the actual disposal site (DIS) (2) the zone around the disposal site (IMZ) and (3) reference

175

samples (REF). An interaction time:SDS was included since time trends could be different at disposal sites versus reference zone. When interaction was significant, the model was applied on each zone separately. Time2 was included to identify non-linear time trends. Outliers were removed from the model by a sequential process when Cooks distance was larger than 0.2. Some extreme values occurred, especially in reference zones, which were not continued in time. The amount of data points included as well as removed outliers are given in Table 1. Changes in proportion of individual PCB and PAH between 2005 and 2006 or 2008e2009 and 2013e2014 at specific dredged spoil disposal sites were analysed by Student t-tests within Statistica 12. 3. Results 3.1. Monitoring of dredged spoil disposal sites The monitoring data for the five dredged spoil disposal sites on the BPNS were modelled. Results are summarized in Table 1 and full models are presented in Table S5. Plots of all time trends, including the intercept, time and time2 coefficients are presented in Fig. S1 in the supplementary data. Focussing on PCB concentrations, none of the dredged spoil disposal sites reveal a significant difference at DIS and IMZ, compared to REF. The analysis of PCB data from 2005 to 2014 revealed no significant trend on dredged spoil disposal sites BR&WS1, BR&WS2 and LNP with average PCB concentrations of 5.12 ± 3.92, 6.07 ± 3.17 and 6.66 ± 3.11 mg kg1 dry weight (d.w.) sediment, respectively. At BR&WZE and BR&WOO, a significant increase could be found at DIS of respectively 36% and 19%, comparing 2005e2006 to 2013e2014 data. The proportion of individual PCB are given in Fig. S2, comparing 2005e2006 with 2013e2014 data. Patterns were similar on each dredged spoil disposal site, with CB153 as main PCB (25.2e30.0%), followed by CB118 (13.8e18.8%), CB138 (14.3e18.6%), CB101 (11.8e18.1%), CB180 (6.9e10.5%), CB52 (5.9e15.5%) and CB28 (4.4e10.1%). A significant increase of CB28 (BR&WS2, BR&WZE, BR&WOO, LNP) and CB101 ratios (BR&WS1, BR&WZE, BR&WOO) within time was compensated by decreased CB156 (BR&WS2, BR&WZE, BR&WOO) and CB52 ratios (BR&WZE, BR&WOO). In contrast to PCB, none of the dredged spoil disposal sites revealed an increase in PAH: whereas no trend was found at BR&WS2, BR&WOO and LNP, a decrease of respectively 14% and 12% can be observed at BR&WS1 and BR&WZE, when comparing data of 2008e2009 to 2013e2014. At BR&WZE, the time trend at the dredged spoil disposal site (DIS) is different from the reference zone: in 2008e2010, the concentration is higher in DIS, decreasing to the level of the reference zone in 2013e2014, while the reference zone did not reveal a significant time trend. Remarkably, the average PAH concentration at BR&WS1 (331 ± 175 mg.kg1 d.w.) was significantly lower than at other dredged spoil disposal sites, with average concentrations varying from 550 ± 111 mg.kg1 d.w. (LNP) to 653 ± 227 mg.kg1 d.w. (BR&WS2). Concentrations at DIS are significantly lower than in the reference zone. Proportions of 3-, 4-, 5- and 6-ring PAH are given in Fig. S3, with maximum percentages at dredged spoil disposal areas of 4-ring PAH (38.2e41.8%). From 2008 to 2009 to 2013e2014, a significant decrease of the 6-ring PAH ratio of 15e21% was noticed, compensated by a small increases of other PAH-fractions. Concentrations of Cd, Pb and Cr are decreasing at BR&WS1, BR&WZE, BR&WOO and LNP. Decreases vary from 10% (BR&WOO) to 37% (BR&WS1) for Pb/Al ratios, 16% (BR&WOO) to 51% (BR&WS1) for Cd/AL ratios and 17% (BR&WZE) to 28% (BR&WOO) for Cr/Al ratios when comparing 2005e2006 to 2013e2014. No decrease is

176

B. De Witte et al. / Chemosphere 156 (2016) 172e180

Table 1 Summary of the monitoring of chemical data on dredged spoil disposal sites. Zone refers to the area for which the model is valid. ALL reflects a model applicable to all zones.

PCB Br&WS1 Br&WS2 Br&WZE Br&WOO LNP PAH Br&WS1 Br&WS2 Br&WZE Br&WZE Br&WZE Br&WOO LNP Al Br&WS1 Br&WS2 Br&WS2 Br&WS2 Br&WZE Br&WZE Br&WZE Br&WOO LNP Fe Br&WS1 Br&WS2 Br&WZE Br&WZE Br&WZE Br&WOO Br&WOO Br&WOO LNP Cd Br&WS1 Br&WS2 Br&WZE Br&WZE Br&WZE Br&WOO LNP LNP LNP Cr Br&WS1 Br&WS2 Br&WZE Br&WZE Br&WZE Br&WOO LNP Cu Br&WS1 Br&WS2 Br&WZE Br&WOO Br&WOO Br&WOO LNP Hg Br&WS1 Br&WS1 Br&WS1 Br&WS2 Br&WZE Br&WZE Br&WZE Br&WOO LNP Ni Br&WS1 Br&WS2 Br&WZE

Zone

Data points

Removed outliers

Time effect

Time trend remark

ALL ALL ALL ALL ALL

223 105 193 178 145

2 e e 2 e

No No Yes Yes No

ALL ALL REF DIS IMZ ALL ALL

127 53 36 61 28 118 97

1 3 1 e e 2 1

Yes No No Yes No No No

Decrease

ALL REF DIS IMZ REF DIS IMZ ALL ALL

237 52 74 22 67 94 42 181 138

e e e 1 e e e 1 2

No No No No No No No No No

Decrease Decrease Decrease Decrease Decrease Decrease Decrease Decrease Decrease

ALL ALL REF DIS IMZ REF DIS IMZ ALL

237 148 67 94 42 54 87 42 140

e 1 e e e e e e e

Yes Yes Yes No Yes Yes No No Yes

Increase which is flattening Increase Increase

ALL ALL REF DIS IMZ ALL REF DIS IMZ

227 147 66 94 42 178 43 69 26

e 1 e e e 1 1 1 e

Yes No No Yes Yes Yes No Yes Yes

ALL ALL REF DIS IMZ ALL ALL

235 149 67 94 42 182 138

2 e e e e e 2

Yes No No Yes Yes Yes Yes

Decrease which is flattening

ALL ALL ALL REF DIS IMZ ALL

237 148 202 54 87 42 140

e 1 1 e e e e

Yes No Yes No Yes No Yes

Decrease which is flattening

REF DIS IMZ ALL REF DIS IMZ ALL ALL

60 149 28 136 67 94 42 178 139

e e e e e e e 4 e

No Yes No Yes No Yes Yes Yes Yes

ALL ALL ALL

236 149 203

1 e e

No No No

Increase followed by slight decrease Increase

SDS effect

Season effect

No No No No No

No No No Higher in spring No

Zone 1 lower No

No No No No No No No

Decrease No No followed followed followed followed followed followed followed followed followed

by by by by by by by by by

increase increase increase increase increase increase increase increase increase

Zone 1 higher

No No No No

Decrease Increase which is flattening

Increase which is flattening

No

Decrease which is flattening

No No

Decrease Decrease followed by increase Decrease which is flattening

No

Decrease which is flattening Decrease which is flattening

Decrease which is flattening Decrease which is flattening Decrease which is flattening Decrease

Decrease which is flattening

Zone 2 lower No

No No No No No

Increase Decrease which is flattening

Zone 1 higher

Increase followed by decrease

in spring in spring in spring in in in in

spring spring spring spring

No No No No Lower in spring No No No No No No No No No Lower in spring No No Lower in spring No No No Higher in spring No No No Lower in spring No No No No No Lower in spring

No Zone 1 higher

No No No Lower in spring No No Higher in spring No Higher in spring

Zone 2 lower No No

Lower in spring No No

Decrease Increase Decrease followed by increase Decrease Decrease which is flattening Decrease which is flattening Decrease

Lower No Lower Lower No Lower Lower Lower Lower

No

B. De Witte et al. / Chemosphere 156 (2016) 172e180

177

Table 1 (continued )

Br&WOO LNP Pb Br&WS1 Br&WS2 Br&WZE Br&WOO LNP Zn Br&WS1 Br&WS1 Br&WS1 Br&WS2 Br&WSE Br&WOO LNP LNP LNP

Zone

Data points

Removed outliers

Time effect

Time trend remark

SDS effect

Season effect

ALL ALL

182 139

e 1

No Yes

Decrease

No No

Lower in spring Lower in spring

ALL ALL ALL ALL ALL

237 148 203 181 140

e 1 e 1 e

Yes Yes Yes Yes Yes

Decrease Decrease Decrease Decrease Decrease

Zone 1 lower No No No

No No No No Lower in spring

REF DIS IMZ ALL ALL ALL REF DIS IMZ

60 149 28 149 203 179 44 70 26

e e e e e 3 e e e

Yes No No No No Yes Yes Yes No

Increase

found at DIS of BR&WS2, even for Pb where the complete model including IMZ and REF reveals a decreasing trend. Hg and Cu concentrations reveal a decreasing trend at most dredged spoil disposal sites. However, for each of these elements, an opposite trend could also be noticed. Hg/Al ratios decrease at BR&WS1, BR&WZE, BR&WOO and LNP, varying from 11% (BR&WOO) to 45% (BR&WS1) whereas an increase of 62.% is noticed at BR&WS2. Cu/Al concentrations decrease at BR&WS1 and LNP with 18 and 28%, respectively. At BR&WOO and BR&WS2, however, there was an increase of Cu/Al ratios of 17% and 28%, respectively, between 2006 and 2014. Ni concentrations were constant at every dredged spoil disposal site, except for a decrease at LNP. Ni/Al ratio decreased with 33% from 2005 to 2006 to 2013e2014. Zn concentrations were not decreasing at BR&WS1, BR&WS2 and BR&WZE while a significant increase in Zn/Al ratios was found at BR&WOO (29%) and LNP (18%) from 2005 to 2006 to 2013e2014. Al and Fe are sometimes used to normalise heavy metal data as they occur naturally in clay minerals (Garcia and Garcia-Gomez, 2005). Al revealed a downward trend followed by an increase on all dredged spoil disposal sites, with no clear net effect between 20052006 and 2013e2014. In the same time period, Fe concentrations increased at BR&WS1 and LNP with 45% and 39%, respectively, while no clear trend was visible at DIS of other dredged spoil disposal sites. In contrast to organic contaminants, seasonal effects were identified within heavy metal time trends. Al, Fe, Cd, Ni, Zn, Pb and Cu reveal lower values in spring at one or more dredged spoil disposal sites whereas results for Hg are mixed and Cr revealed a higher value in spring only at BR&WZE. In particular at LNP, most heavy metals showed lower concentrations in spring compared to autumn. 3.2. Assessment according to OSPAR environmental criteria For marine environmental monitoring, the OSPAR commission has determined background assessment concentrations (BAC) and EAC. BAC values indicate whether contamination levels are “near background” (for naturally occurring substances) or “close to zero” (for man-made substances). EAC values represent the contaminant concentration in the environment below which no chronic effects are expected to occur in marine species, including the most sensitive species (OSPAR, 2009b). Values are normalised to 2.5% total organic carbon (PAH, PCB) or 5% Al (Hg, Pb, Cd) in order to compare with BAC and EAC values. For PAH, Cd, Hg and Pb, no EAC values

which is flattening which is flattening which is flattening which is flattening

Increase which is flattening Increase Increase

No Zone 1 lower No

No Lower in spring No No No No Lower in spring No No

were proposed. Effect Range Low (ERL) values, developed by USEPA, are given, not normalised to TOC or Al. ERL values also indicate the concentration below which effects are not likely (OSPAR, 2009b). The ER-Low (ERL) value is defined as the lower tenth percentile of the data set of concentrations in sediments which were associated with biological effects. Adverse effects on organisms are rarely observed when concentrations fall below the ERL value, and the ERL therefore has some parallels with the philosophy underlying the OSPAR EACs and WFD EQSs. The procedure by which ERL criteria are derived is very different from the methods of derivation of EACs and EQSs, and so precise equivalence between the two sets of criteria should not be expected. ERL values are to be used in sediment assessments of contaminants (e.g. PAH and metals) as an interim solution where recommended EACs are not available. For each dredged spoil disposal site, average and normalised DIS data of 2013e2014 are presented in Table 2, indicating the current status of the disposal site. Cd and Pb values are near BAC values. Hg, PCB and most PAH values are above their BAC values. CB118 is even above the EAC value at all dredged spoil disposal sites. ERL values of Hg, Cd and Pb and PAH values are difficult to compare since no normalisation is applied to ERL values and analysis was done on the <63 mm fraction. However, given the fact that contaminants occur at higher concentrations in the fine fraction, whole sample analysis will result in lower concentrations, most probably not exceeding ERL values. 3.3. Monitoring at ports Complete results of harbour monitoring are given in Table S6 and Table S7. At Fig. 2, a selection of results is presented, normalised to Al for heavy metal data and to TOC for PAH and PCB data. These values are compared to the average of normalised DIS and REF data from 2013 to 2014 in which REF-data is scaled to 100. For all compounds measured, concentration differences are limited between Nieuwpoort harbour sampling points. For the 2013 sampling, relative standard deviation (RSD) varied from 6.5% to 26.7%, except for Zn with 63.6% RSD while the RSD for the 2014 sampling varied from 5.7% to 23.1%. At harbour Oostende, much higher differences can be noted. At sampling locations HOO01, HOO02, HOO03, HOO05, HOO06 and HOO11, concentrations of PCB, PAH, Cd, Pb, Cr, Cu and Hg are clearly elevated compared to Nieuwpoort harbour, but differences are less than a factor 2. For these locations, HOO06 is selected as

178

B. De Witte et al. / Chemosphere 156 (2016) 172e180

Table 2 Comparison with EAC (PCB) and ERL (PAH and metals) data of DIS 2013e2014 data. EAC values are given in bold. Cd and Pb values are reported in mg.kg1 d.w.. All other values are reported in mg.kg1 d.w. All data are normalised to 2.5% TOC (PAH, PCB) or 5% Al (Hg, Cd, Pb). BAC values are normalised to 2.5% TOC or 5% Al, EAC values are normalised to 2.5% TOC, ERL values are not normalised. BR&WS1 Naphthalene Phenanthrene Anthracene Fluoranthene Pyrene Benzo[a]anthracene Chrysene Benzo[a]pyrene Benzo[ghi]perylene Indeno[123cd]pyrene CB28 CB52 CB101 CB118 CB138 CB153 CB180 Hg Cd Pb a

22.8 33.6 13.5 64.0 48.3 27.4 25.1 32.1 29.7 34.7 0.39 0.48 1.03 0.97 0.89 1.41 0.46 66.1 0.15 22.8

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

8.9 12.8 5.5 24.6 17.9 10.7 9.3 13.3 12.2 14.5 0.16 0.25 0.36 0.24a 0.21 0.34 0.25 33.1 0.06 5.1

BR&WS2

BR&WZE

BR&WOO

LNP

BAC

EAC/ERL

54.3 ± 17.1 72.5 ± 21.7 25.5 ± 8.0 132.9 ± 37.3 98.8 ± 28.5 58.8 ± 17.5 52.8 ± 17.1 68.2 ± 23.4 61.9 ± 16.3 69.0 ± 18.0 0.78 ± 0.18 0.43 ± 0.20 1.15 ± 0.27 1.41 ± 0.38a 1.32 ± 0.30 2.01 ± 0.36 0.59 ± 0.16 178.0 ± 29.7 0.37 ± 0.08 37.4 ± 5.1

41.9 ± 9.2 56.7 ± 11.0 23.0 ± 5.2 112.8 ± 26.7 81.4 ± 19.4 48.3 ± 11.4 38.8 ± 8.5 56.0 ± 13.0 52.4 ± 11.0 64.0 ± 16.7 0.72 ± 0.15 0.54 ± 0.15 1.07 ± 0.23 1.22 ± 0.26a 1.17 ± 0.24 2.03 ± 0.43 0.80 ± 0.25 176.3 ± 36.9 0.37 ± 0.05 36.0 ± 4.6

40.0 ± 12.2 56.8 ± 11.0 21.6 ± 3.8 111.3 ± 25.6 80.8 ± 19.8 48.2 ± 10.0 41.7 ± 7.3 56.8 ± 11.6 54.3 ± 7.9 63.3 ± 14.4 0.64 ± 0.22 0.48 ± 0.09 1.03 ± 0.25 1.22 ± 0.37a 1.14 ± 0.38 1.84 ± 0.48 0.64 ± 0.19 175.9 ± 22.8 0.38 ± 0.06 37.6 ± 4.1

37.6 ± 15.1 47.9 ± 15.4 19.7 ± 7.0 99.0 ± 42.6 73.7 ± 32.7 45.5 ± 19.9 37.2 ± 13.7 55.8 ± 21.1 52.5 ± 15.1 63.3 ± 24.2 0.57 ± 0.22 0.43 ± 0.07 0.98 ± 0.24 1.26 ± 0.34a 1.23 ± 0.27 1.77 ± 0.52 0.48 ± 0.22 167.1 ± 23.2 0.35 ± 0.06 41.7 ± 4.3

8 32 5 39 24 16 20 30 80 103 0.22 0.12 0.14 0.17 0.15 0.19 0.10 70 0.31 38

160 240 85 600 665 261 384 430 85 240 1.7 2.7 3.0 0.6 7.9 40 12 150 1.2 47

Exceedance of EAC.

20 m was done, with sampling in 2013 just in front of a workplace for leisure boat maintenance where no dredging could take place towards sampling within the nearby frequently dredged zone in 2014. HOO10 revealed much lower concentrations of PCB, PAH and Hg compared to other harbour Oostende as well as harbour Nieuwpoort sampling locations. HOO10 is located upstream on the channel that connects harbour Oostende with the hinterland. Compared to DIS and REF-data (Fig. 2), HOO04, HOO07 and HOO08 reveal increased concentrations of several organic and inorganic contaminants, with maximum concentrations relatively compared to REF at HOO08 of 1755% (Cu/Al), 458% (PCB/TOC) 447% (Zn/Al), 193% (PAK/TOC), 182% (Cd/Al), 148% (Hg/Al) and 125% (Pb/ Al). In contrast, harbour Nieuwpoort sampling points do not reveal an enrichment, since each pollutant has concentrations of 68%e 105% compared to normalised REF-data. 4. Discussion Fig. 2. Concentration of PCB, PAH and heavy metals at different Harbour Oostende sampling points, the average of harbour Nieuwpoort sampling points and DIS sampling points (2013e2014), relative to REF-sampling points (2013e2014). Heavy metal data is normalised to Al, PCB and PAH data to TOC. HOO04 and HOO08 reveal increased pollution, HOO10 decreased pollution, HOO06 is representative for most other HOO sampling points.

representative in Fig. 2. At HOO04 and HOO08, very high concentrations of Cu, Zn and PCB are found, and PAH concentrations are also elevated compared to other harbour Oostende locations. Maximum concentrations reached 322 mg.kg1 d.w. for Cu, 789 mg.kg1 d.w. for Zn and 53.9 mg.kg1 d.w. for PCB at HOO08. HOO04 and HOO08 are both located at the same dock. At this site of the harbour, there has been no dredging activity for many years. At HOO07, a high pollution grade was notified in 2013, with a Cu concentration of 1641 mg.kg1 d.w., a Zn concentration of 1212 mg.kg1 d.w. and a PCB concentration of 15.8 mg.kg1 d.w. In 2014, concentrations were much lower, but still clearly elevated compared to low polluted Oostende harbour locations HOO01-03, HOO05-06 and HOO11. It is important to note that this is the only sampling point where 2013 and 2014 sampling was not done on exactly the same place. A migration of sampling location of about

4.1. Trends in organic contaminant concentrations The use of PCB is banned since mid-80s (OSPAR, 2012). Nevertheless, PCB levels are still clearly elevated compared to background assessment criteria and proposed EAC values are exceeded for CB118 at dredged spoil disposal sites as well as corresponding reference zones. Moreover, no decrease of PCB concentrations was found between 2005 and 2014, it even slightly increased at two disposal sites (BR&WZE, BR&WOO). This was already noted by Roose et al. (2005) for the period 1991e2001 and the OSPAR Quality Status Report (OSPAR, 2010a) for the partly overlapping period 1998e2007. These findings do not cope with the model of Everaert et al. (2014) who found 50e66% decrease in a data series from 1991 till 2010. It is not entirely clear to the authors why this last model predicts a different trend. Major differences with our own assessment seems to be the longer timeframe and the use of data from different data owners, implying differences in methodology and quality control. Several factors influence the environmental concentrations of PCB: inputs from rivers and dredged material (Laane et al., 1999), atmospheric deposition, degradation, volatilisation and transport with currents (O’Driscroll et al., 2013). The PCB-ban definitely

B. De Witte et al. / Chemosphere 156 (2016) 172e180

caused a rapid decline in PCB concentrations in the ‘80s and the beginning of the ‘90s which levelled thereafter. Sobek et al. (2015) demonstrated this scenario in sediment from the Baltic Sea and Laane et al. (1999) measured a decrease up to 80% in 1981 till 1996 in sediment of the Dutch coastal zone, adjacent the BPNS. The combination of our data and those of Roose et al. (2005) and the OSPAR Quality Status report (OSPAR, 2010a), suggest that the inputs from rivers, atmospheric deposition and dredged spoil disposal are of the same order of magnitude as the dilution, outputs and losses during the last 10e15 years. This is strengthened by the measurement of high PCB concentration at harbour Oostende sampling points (HOO04, HOO07, HOO08) which shows the importance of harbour dredged spoil disposal as possible input source of PCB to the marine environment. PAH concentrations are higher than background assessment criteria. This is not surprising given the high degree of industrialisation of the area in this part of Western Europe and the intense marine traffic in the BPNS. However, comparison of monitoring data with ERL values indicate that PAH associated environmental risks are low. PAH concentrations revealed a downward trend at disposal sites Br&WS1 and BR&WZE, although these disposal sites receive dredged spoil form the industrialized harbour Zeebrugge and the river Scheldt. Since no clear trend in origin of the dredged spoil can be seen, this reflects a remediating effect of targeted emission reductions by the industry and maritime activity. In future, however, it is not expected to have a further decline of PAH concentrations as increased emissions from outside the OSPAR convention area and increased energy consumption will probably counteract point source reduction measures (OSPAR, 2010b; Gao et al., 2013). 4.2. Trends in heavy metal concentrations Since ERL values at the BPNS are not exceeded for Hg, Cd and Pb, current environmental risks for heavy metals are expected to be low. Trends of heavy metal concentrations are generally in line with OSPAR observations. At most dredged spoil disposal sites, a decrease of Cu, Cd, Pb, Cr and Hg can be related to stringent pollution control measures (OSPAR, 2010a). At Br&WS2, however, heavy metal trends are not in line with other dredged spoil disposal sites while Hg concentrations are even increasing. At 2005e2006, Hg levels at BR&WS2 were low (99.9 mg.kg1 d.w.), increasing to 147 mg.kg1 d.w. in 2013e2014. At this remote site, the yearly dredged spoil disposal rate strongly increased from 2009 to 2014 compared to 2005e2008 without clear difference in origin of the dredged spoil (Table S1). It is suggested that increased dredged spoil disposal rates resulted in higher contaminant loads which opposed natural trends. Although Hg levels are still lower than ERL values and even lower than Hg levels of BR&WZE (163 mg.kg1 d.w.), this increasing trend should be conscientiously followed in time. Again, this clearly indicates the importance of routinely monitoring the chemical quality at dredged spoil disposal sites within the MSFD framework. Upwards trends in Zn at disposal sites BR&WOO and LNP and Cu at disposal site BR&WOO are found, despite the least intensive use of these sites in comparison with the other sites and the low impacted dredged spoil they receive. After the ban on Pb and triorganotin, Cu and Zn based paints are dominating the marine antifouling market (Turner, 2010). High concentrations of Cu and Zn are reported in harbours nearby boat- and shipyards. These compounds may enter the marine environment through leaching or by the release of fine paint particles during blasting of boat hulls (Berto et al., 2012; Costa and Wallner-Kersanach, 2013; Singh and Turner, 2009; Turner, 2010; Ytreberg et al., 2010). Ytreberg et al. (2010) report substantially higher Zn concentrations in leisure

179

boat paints compared to paints for professional ship hulls. Since Br&WOO & LNP both contain a relative large leisure boat harbour, a possible link was suggested. To investigate a relationship between Zn and/or Cu contamination with antifouling agents and time trends at dredged spoil disposal sites, additional harbour sampling was performed. At Nieuwpoort harbour, four sampling points were nearby a shipyard (HNP06eHNP09). However, none of these locations revealed increased Cu and Zn concentrations. At Nieuwpoort, boatyards have adapted strict environmental measures between 2000 and 2010, catching rinsing water and paint spills. In combination with frequent dredging, it can be concluded that Cu and Zn sources are limited and that pollution in the upper sediment layers is already dredged to the dredged spoil disposal site. In contrast, HOO04 and HOO08 at Oostende harbour reveal increased concentrations for both harbour sampling campaigns, HOO07 for one harbour sampling campaign. HOO04 and HOO08 are located inside a dock, which is not routinely dredged. It contains an active shipyard, the remains of a historic shipyard and a small metallurgical company. HOO07 is nearby an active shipyard. Whereas most heavy metal concentrations show a slight increase at HOO04, HOO07 and HOO08 compared to other harbour sampling points, concentrations of PCB, Zn and Cu are extremely high, clearly linked to these sources. Due to remobilisation of heavy metals, diffusion, turbidity and tidal working, this will also affect the dredged zones of Oostende harbour and its disposal site to some extent. This was shown at HOO07, where Cu and Zn concentrations at 2014 sampling were also elevated in the dredged zone of the harbour. The size of this harbour effect is unclear and it is, by consequence, not sure this may lead to a significant increase of Cu and Zn concentrations at BR&WOO. Strict monitoring should be continued to follow up these trends and environmental assessment criteria should be established for these compounds to assess environmental quality for Cu and Zn within MSFD. If shipyards play an important role in the level of Cu and Zn concentrations at dredged spoil disposal sites, concentrations at BR&WOO are expected to further increase while a decrease can be expected at LNP due to the pollution control measures. It is also advisable that catching rinsing water and paint spills become standard at boat- and shipyards, not only in the studied harbours. Seasonal changes in heavy metal concentrations are affected by several factors, and since this was not the main goal of this study, metadata to assess this seasonal affects such as redox potential, pH values and sulfide or oxygen concentration are not measured. Goa et al. (2009) suggested that metal concentration at the Belgian coast in March are especially driven by low dissolved sulfide concentrations which may explain lower concentration in the late winter e early spring sampling campaign. Since sampling campaigns are shortly before or 3 months after the spring algal bloom (Gao et al., 2009), the influence of organic material is suspected to be low in the trends presented. 5. Conclusions Monitoring PCB, PAH and heavy metals is part of the MSFD monitoring scheme. Comparing current concentrations of these compounds at dredged spoil disposal sites at the BPNS with EAC and ERL data revealed that the risk of chronic effects by dredged spoil disposal at the BPNS is limited or not higher compared to reference zones. However, the lack of decrease of PCB values, the increase of Zn and Cu values at the dredged spoil disposal sites LNP and/or BR&WOO and the systematic increase of Hg values at BR&WS2 stress the importance to continue trend monitoring, which also gives the opportunity to counteract negative trends by adapting dredged spoil disposal rates, closing dredged spoil

180

B. De Witte et al. / Chemosphere 156 (2016) 172e180

disposal sites or taking pollution control measures at the source. Increase of Zn and Cu concentrations was suggested to be affected by boatyards at the harbours. Although the extend was not quantified, it is advisable to take pollution control measures at every boat- and shipyard e.g. by catching rinsing water and paint spills. Acknowledgements We acknowledge financial support from the Flemish department of mobility and public works through the project “bagger”. Shiptime on RV Belgica was financed by Belspo and RBINS OD Nature. Harbour sampling was made possible by the support and deric Van Steen, infrastructure of VLIZ. We also want to thank Fre Kevin Vanhalst and the laboratory technicians of the ILVO chromatography lab for technical support. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.chemosphere.2016.04.124. References Anon. (2015a). Port of Zeebrugge, jaarverslag 2014, 24. Anon. (2015b). Port of Antwerp, annual report 2014, 69. Anon. (2015c). Haven Oostende, jaarverslag 2014, 95. Baeyens, W., Leermakers, M., De Gieter, M., Nguyen, H.L., Parmentier, K., Panutrakul, S., Elskens, M., 2005. Overview of trace metal contamination in the Scheldt estuary and effect of regulatory measures. Hydrobiologia 540, 141e154. Berto, D., Brusa, R.B., Cacciatore, F., Covelli, S., Rampazzo, F., Giovanardi, O., Giani, M., 2012. Tin free antifouling paints as potential contamination source of metals in sediments and gastropods of the southern Venice lagoon. Cont. Shelf Res. 45, 34e41. Costa, L.D.F., Wallner-Kersanach, M., 2013. Assessment of the labile fractions of copper and zinc in marinas and port areas in Southern Brazil. Environ. Monit. Assess. 185, 6767e6781. Donata, L., 2010. Polycyclic Aromatic Hydrocarbons (PAHs) Factsheet. JRC 60146, third ed. Institute for Reference Materials and Measurements, p. 25. Everaert, G., De Laender, F., Deneudt, K., Roose, P., Mees, J., Goethals, P.L.M., Janssen, C.R., 2014. Additive modelling reveals spatiotemporal PCBs trends in

marine sediments. Mar. Pollut. Bull. 79, 47e53. Garcia, J.M., Garcia-Gomez, J.C., 2005. Assessing pollution levels in sediments of a harbour with two opposing entrances. Environmental implications. J. Environ. Manage. 77, 1e11. Gao, Y., Lesven, L., Gillan, D., Sabbe, K., Billon, G., De Galan, M.E., Baeyens, W., Leermakers, M., 2009. Geochemical behavior of trace elements in sub-tidal marine sediments of the Belgian coast. Mar. Chem. 117, 88e96. Gao, Y., De Brauwere, A., Elskens, M., Croes, K., Baeyens, W., Leermakers, M., 2013. Evolution of trace metal and organic pollutant concentrations in the Scheldt River Basin and the Belgian Coastal Zone over the last three decades. J. Mar. Syst. 128, 52e61. Heim, S., Schwarzbauer, 2013. Pollution history by sedimentary records: a review. Environ. Chem. Lett. 11, 255e270. Laane, R.W.P.M., Sonneveldt, H.L.A., Van der Weyden, A.J., Loch, J.P.G., Groeneveld, G., 1999. Trends in the spatial and temporal distribution of metals (Cd, Cu, Zn and Pb) and organic compounds (PCBs and PAHs) in Dutch coastal zone sediments from 1981 to 1996: a model case study for Cd and PCBs. J. Sea Res. 41, 1e17. Lauwaert, B., Fettweis, M., De Witte, B., Devriese, L., Van Hoey, G., Timmermans, S., Martens, C., 2014. Vooruitgangsrapport (juni 2014) over de effecten op het mariene milieu van baggerspeciesstortingen, p. 24. OD Nature report vol. BL/ 2014/01. O’Driscroll, K., Mayer, B., Ilyina, T., Pohlmann, T., 2013. Modelling the cycling of persistent organic pollutants (POPs) in the North Sea system: fluxes, loading, seasonality, trends. J. Mar. Syst. 111e112, 69e82. OSPAR, 2009a. OSPAR Guidelines for the Management of Dredged Materials. OSPAR Commision reference 2009e4. OSPAR, 2009b. Agreement on CEMP assessment criteria for the QSR 2010, agreement number 2009-2, publication number 2009/461, 7p. OSPAR, 2010a. Quality Status Report 2010. OSPAR Commission, London, p. 176. OSPAR, 2010b. OSPAR Coordinated Environmental Monitoring Programme (CEMP). OSPAR commission, London. Agreement 2010e1. OSPAR, 2012. CEMP Assessment Report. Monitoring and Assessment Series. OSPAR Publication, 563/2012. ISBN 978-1-907390-68-5. Roose, P., Raemaekers, M., Cooreman, K., Brinkman, U.A.T., 2005. Polychlorinated biphenyls in marine sediments from the southern North Sea and Scheldt estuary: a ten-year study of concentrations, patterns and trends. J. Environ. Monit. 7, 701e709. Sobek, A., Sundqvist, K.L., Assefa, A.T., Wiberg, K., 2015. Baltic Sea sediment records: unlikely near-future declines in PCBs and HCB. Sci. Total Environ. 518e519, 8e15. Singh, N., Turner, A., 2009. Leaching of copper and zinc from spent antifouing paint particles. Environ. Pollut. 157, 371e376. Turner, A., 2010. Marine pollution from antifouling paint particles. Mar. Pollut. Bull. 60, 159e171. Ytreberg, E., Karlsson, J., Eklund, B., 2010. Comparison of toxicity and release rates of Cu and Zn from anti-fouling paints leached in natural and artificial brackish seawater. Sci. Total Environ. 408, 2459e2466.