Contemporary ecological threats from historical pollution sources: impacts of large-scale resuspension of contaminated sediments on sessile invertebrate recruitment

Authors

  • Nathan A. Knott,

    Corresponding author
    1. Evolution and Ecology Research Centre, School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, NSW 2052, Australia
      *Corresponding author. E-mail: nathan.knott@environment.nsw.gov.au
    Search for more papers by this author
  • Joel P. Aulbury,

    1. Evolution and Ecology Research Centre, School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, NSW 2052, Australia
    Search for more papers by this author
  • Trevor H. Brown,

    1. Port Kembla Port Corporation, Military Rd, Port Kembla, NSW 2505, Australia
    Search for more papers by this author
  • Emma L. Johnston

    1. Evolution and Ecology Research Centre, School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, NSW 2052, Australia
    Search for more papers by this author

*Corresponding author. E-mail: nathan.knott@environment.nsw.gov.au

Summary

  • 1 Historical contamination continues to pose a serious ecological threat to many ecosystems across the world. Marine sediments have long acted as sinks for contaminants from surrounding industry and urbanization, and sediment-bound contaminants are known to affect the ecology of sediment infauna. When sediments are disturbed, however, contaminants are resuspended and potentially released into the water-column and dispersed to other environments. The threat posed by the resuspension of contaminated sediments has been a focus of substantial research in the geochemical and ecotoxicology fields, yet to date there has been no ecological assessment of the impacts of a real-world resuspension event involving contaminated sediments.
  • 2 We assessed the ecological threat posed by the resuspension of contaminated sediments by testing for impacts of a major dredging operation in an estuary with highly contaminated sediments. We sampled the recruitment of sessile invertebrates in this estuary and two external reference estuaries, before and during dredging. These invertebrates are filter-feeders that settle and live on hard substrata above the seafloor. Impacts of the dredging-related resuspension were tested using a Beyond BACI analysis.
  • 3 Dredging activities resulted in the large-scale resuspension of contaminated sediments. Concurrently, the recruitment of the dominant filter-feeders (e.g. barnacles and polychaete worms) was virtually extinguished for 4 months, despite being abundant prior to dredging. This pattern contrasted with the recruitment of the same invertebrates in the reference estuaries, which showed little change over the same period.
  • 4Synthesis and applications. The severe decrease in the recruitment of sessile invertebrates within an estuary exposed to dredging and the deposition of contaminated sediments indicates that the resuspension of these sediments pose a real ecological threat to organisms in contact with the contaminated water-column. Containment measures (e.g. silt curtains) are, therefore, essential and further engineering innovations (e.g. dredging designs and operations) are necessary to reduce resuspension during the dredging and deposition of sediments. In addition, this study demonstrates that past pollution events can cause current ecological impacts that extend well beyond those habitats recognized as being contaminated.

Introduction

Despite the dramatic improvement in the management and disposal of contaminants in many parts of the world (Valette-Silver 1993; Hornberger et al. 2000), historical contamination poses a real threat to the current ecology of many systems (Sheppard 2005; Norris et al. 2007; Johnston & Roberts 2009). There is substantial knowledge and understanding of the direct impacts on the ecology of organisms inhabiting contaminated soils and sediments (Millward et al. 2004; Ramsey et al. 2005). The threat of these contaminants may, however, extend well beyond those habitats if contaminants can move to other environments (Larsson 1985; Asare et al. 2000; Coulthard & Macklin 2003).

In the marine environment, harbours worldwide have long histories of acting as sinks for contaminants from surrounding industry and urbanization (Cundy et al. 2003; Taylor, Birch, & Links 2004). The major contaminants that flow into harbours are metals, tributyltins, polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyls (PCBs) and pesticides (Fowler 1990). Many of these contaminants rapidly bind to particulate matter in the water-column (McLusky, Bryant, & Campbell 1986) and eventually sink to the seafloor (Birch & Taylor 1999; Cundy et al. 2003). Bound to the sediment, the ecological effects of contaminants are thought to be largely contained (McLusky et al. 1986; Stauber et al. 2000) and mainly affect fauna directly associated with the sediments (i.e. infaunal invertebrates; Clements 2004; Trannum et al. 2004).

The disturbance, or resuspension, of contaminated sediments is a mechanism by which the ecological threat can be transferred beyond the seafloor and potentially affect organisms in the water-column (Birch 2000; Simpson, Apte, & Batley 1998). Sediments have the potential to be remobilized by a range of natural and anthropogenic disturbances such as dredging, trawling, storms, tides and bioturbation (Eggleton & Thomas 2004). There is strong evidence that the resuspension process will release contaminants into the surrounding water-column, and that contaminants may then become biologically available (reviewed in Eggleton & Thomas 2004). Furthermore, at high concentrations, there can be direct effects of suspended sediment (Wilber & Clarke 2001), such as damage to the gills and eyes of fish (Johnston 1981) and clogging of the filtering apparatus of invertebrates (Airoldi 2003). Current chemical and ecotoxicological research suggests that the resuspension of contaminated sediments may pose a real and important threat to the ecology of water-column organisms (Munns, Berry, & Dewitt 2002; Nayar, Goh, & Chou 2004), but there remains no assessment of the ecological impacts from a real-world resuspension event (Eggleton & Thomas 2004). This might because of the large spatial scales involved with these disturbances, the difficulties coordinating and timing assessments of large-scale anthropogenic disturbance, and the perceived difficulties with assessing such ecological effects in the field. Nevertheless, this represents a significant knowledge gap, as the resuspension of contaminated sediments occurs frequently in many contaminated estuaries and ports across the world (Johnston 1981; Eggleton & Thomas 2004).

In this study, we utilized a large-scale dredging project in an estuary (Port Kembla, NSW, Australia) with highly contaminated sediments to assess the potential ecological impacts caused by the resuspension of such sediments. We used sessile invertebrates as model organisms to assess the ecological impacts because (i) they are a significant component of the biological diversity of many estuaries; (ii) they are suspension-feeders which are generally sensitive to elevated metal loadings (Hall, Scott, & Killen 1998; Johnston & Keough 2002); and (iii) they cannot move away from the resuspensions. Furthermore, the recruitment of sessile invertebrates also includes the arrival of larvae, their attachment, metamorphosis and early survival. Early life-history stages provide a sensitive test of the effects of contaminants (Connor 1972; King, Gale, & Stauber 2006), and the ongoing development of the assemblage enables assessment of the indirect effects of the disturbance via species interactions (Johnston & Keough 2003).

We used a ‘Beyond BACI’ experimental design (Underwood 1994) to test for immediate impacts of the dredging-related sediment resuspension. This involved sampling the recruitment of invertebrates twice before and twice during dredging within a dredged estuary and also at two other external reference estuaries. This design has been demonstrated to provide robust inference of numerous anthropogenic impacts (Underwood 1994; Downes et al. 2002). In this study, we specifically tested the prediction that the resuspension of contaminated sediments by dredging activities (i.e. dredging and deposition of dredged sediments) would decrease the recruitment of sessile invertebrates within an estuary, relative to external reference estuaries.

Materials and methods

Contamination history and dredging operations in Port Kembla

Port Kembla, Australia, has had heavy industry based along its shoreline for over 80 years, including steel, tin, copper and zinc manufacturing and processing (He & Morrison 2001). A large urban area also covers the catchment that drains into the harbour. The accumulation of contaminants draining into the estuary over many years has led to high concentrations of cadmium, chromium, copper, lead, mercury, nickel and zinc in the sediments as well as PAHs, tributyltin, PCBs and pesticides (Table 1; PBP 2005). Lead and zinc, in particular, are found in concentrations that are five and nine times greater than Australian sediment quality trigger values respectively (Table 1; ANZECC 2000).

Table 1.   Contaminant concentrations in dredged or excavated sediments from the Inner Harbour of Port Kembla Thumbnail image of

The Port Kembla Port Corporation (PKPC) required dredging of extensive amounts of sediments from the harbour and excavation of adjacent land to create extra shipping berths (in total c. 325 000 m3 of sediment). Much of this sediment (59%) was contaminated and was considered unsuitable for disposal at sea (PBP 2005). Highly contaminated estuarine clays (0·5–4 m thick) covered the floor of the harbour (Table 1). These sediments were dredged from three areas within the inner section of the harbour (referred to as the Inner Harbour) using cutter-suction dredging (27 348 m3, Fig. 1) and deposited into a deep water zone within the Inner Harbour (Fig 1). Firmer underlying sediments and land sediments adjacent to the harbour were also excavated and contained moderately contaminated sediments (i.e. 163 761 m3, Table 1). These sediments were deposited into the outer section of the harbour (referred to as the Outer Harbour, Fig. 1). Dredging operations commenced in March 2007 and were completed in October 2007.

Figure 1.

 Sampling areas within the Inner and Outer Harbours of Port Kembla and the reference estuaries of Botany Bay and Port Hacking. Boxes (bsl00001) indicate Inner Harbour and comparison sites and circles (•) indicate Outer Harbour and comparison sites. Also, shown are the dredging (inline image) and deposition (bsl00036) areas within the Inner and Outer Harbours.

The dredging and deposition of extensive volumes of fine contaminated sediments was likely to create a large-scale resuspension event which was of substantial concern to governmental regulatory agencies and the PKPC. Efforts were, therefore, made to reduce the resuspension by using cutter-suction dredging and contained low energy deposition in the Inner Harbour, and a silt curtain in the Outer Harbour. In this study, we provide an assessment of whether the dredging of sediments, even with control measures in place, posed a threat to water-column organisms.

Assessment design

A Beyond BACI design was used to test for potential ecological impacts of the dredging-related resuspension of contaminated sediment in Port Kembla. As dredging activities were occurring in several sections of the Inner Harbour, and also in the Outer Harbour, they were both considered as potentially impacted locations (PIL). Potential changes in the recruitment of sessile invertebrates due to the dredging-associated resuspension of contaminated sediments were tested by comparing changes in recruitment in Port Kembla before and during dredging, relative to changes in recruitment at two other reference locations outside of Port Kembla.

External reference estuaries were Port Hacking and Botany Bay (Fig. 1). Two sections of each of these estuaries were compared with the two sections of Port Kembla (i.e. the Inner and Outer Harbours). These areas were selected on the basis of similarities to the Inner and Outer Harbours in terms of physicochemical water conditions (N.A. Knott, J.P. Aulbury & E.L. Johnston, unpublished data) and distance from the ocean. Recruitment in the Inner Harbour was compared with that in Burraneer Bay in Port Hacking (Fig. 1) and in the San Souci–Sylvania–Blakehurst upstream of Botany Bay (Fig. 1), and recruitment in the Outer Harbour was compared with that between Hungry and Burraneer Points in Port Hacking (Fig. 1) and the northern section of Botany Bay (Fig. 1). The Botany Bay and Port Hacking estuaries cover a wide range of estuary attributes – size, surrounding anthropogenic activities and general physicochemical conditions. Nevertheless, they all had similar compositions of sessile invertebrate species (i.e. presence/absence in the first two rounds of sampling) and thus were appropriate as reference locations (Glasby & Underwood 1998; Downes et al. 2002).

Recruitment sampling

At each round of sampling, recruitment plates were deployed at four to six sites representatively across each sampling area within each estuary. At each site, two recruitment plates (roughened black perspex, 110 × 110 × 4·5 mm) were attached by stainless steel bolts to a backing panel (grey PVC, 300 × 300 mm). These were deployed vertically at a depth of c. 2 m (MLWS) from the surface and between 2 and 5 m from the seafloor. Each panel was hung from an existing artificial structure (e.g. wharf, jetty, pontoon or marker buoy) or, where no such structures existed, deployed attached to a weighted line with a subsurface float. Each set of plates was deployed for 8 weeks to allow an assemblage of sessile invertebrates to develop. Plates were deployed twice ‘before’ and twice ‘during’ the dredging. ‘Before’ dredging sampling times were between mid-October 2006 and mid-February 2007. ‘During’ dredging sampling times were between early March and end of June 2007. There were 112 and 74 samples in total for the Inner and Outer Harbour comparisons respectively.

After each 8-week deployment, the plates were collected and the assemblages were sampled. Sampling involved identifying and counting each invertebrate species and other variables (e.g. bare space) under 100 evenly spaced points (i.e. providing percentage cover data) using a dissecting microscope. Taxa which were on a plate but not under a point were recorded as having a nominal cover of 0·5%. The inside 10 × 10 cm of the plate were sampled to avoid edge effects, and we avoided a 2-cm diameter central area of the plate surrounding the bolt head.

Turbidity

Turbidity measurements were taken to characterize potential increases in turbidity that may have occurred during the dredging operations. Measurements were taken at the recruitment plates several times opportunistically in each estuary prior to and during dredging. Turbidity measurements were made using a Yeokal 611 physicochemical probe (Yeokal Pty Ltd, Brookvale, NSW, Australia) at similar depths as the recruitment plates. Turbidity measurements were collected from Port Kembla and Port Hacking during the ‘before’ period, but technical problems prevented collection of data from Botany Bay throughout this period. Data were collected from each estuary throughout the ‘during’ period.

Aqueous metal concentration

Metal concentrations within Port Kembla are routinely sampled by CSIRO (Centre for Environmental Contaminants Research; CECR) as part of the environmental monitoring within the harbour for BlueScope Steel Pty Ltd. (Port Kembla, NSW, Australia). Samples collected from this separate monitoring programme were used to evaluate whether metals were released because of the dredging-related resuspension. This programme sampled three sites in both the Inner and Outer Harbours quarterly. Two sets of samples were collected during dredging (i.e. during: April and August 2007) and two sets of samples were used for comparison from similar times from the year before dredging (i.e. before: April and June 2006). At each site, 1 L of water was collected in new polyethylene bottles, filtered (0·45 μm) and acidified with nitric acid. Analyses of dissolved metal concentrations were carried out by CSIRO (CECR) with appropriate QA standards (http://www.clw.csiro.au/cecr/). Samples were analysed for 12 metals using an inductively coupled plasma atomic emission spectrometer [metals and detection limits: Al (1 μg L−1), As (0·2 μg L−1), Cd (2 μg L−1), Cr (3 μg L−1), Cu (2 μg L−1), Fe (2 μg L−1), Mn (1 μg L−1), Ni (3 μg L−1), Pb (1 μg L−1), Se (0·1 μg L−1), Sn (1 μg L−1) and Zn(2 μg L−1)], but only four were detected during the study period (i.e. Al, Mn, Se and Zn). Total suspended sediments (TSS) were also sampled in conjunction with the metals.

Statistical analysis

The Beyond BACI design (Underwood 1994) consisted of five factors: (i) ‘Period’ (Pe) was an orthogonal factor with two levels – before and during dredging, (ii) ‘Time’ (Ti) was a random factor nested within ‘Period’ and had two levels, (iii) ‘Potentially impacted location’ vs. ‘Reference locations’ (PIL vs. Refs) was an orthogonal fixed factor with two levels (i.e. PIL and References), (iv) Estuary (E) was a random factor nested within PIL vs. Refs and had either one (PIL) or two levels (Refs), and (v) ‘Site’ (Si) was a random factor nested within ‘Estuary’ and there were two replicate samples within each site (n = 2). The numbers of sites sampled within each estuary and time varied between 3 and 4 due to the random loss of samples throughout the study. Separate tests were carried out to assess impacts in the Inner and Outer Harbours (N = 112 and 74). As the analysis compares whether the mean of the PIL is within the range of variation observed at the references the PIL is presented as a mean without any SE (because there was only one potentially impacted estuary per analysis) and the references estuaries are presented as a mean with a SE (because there were two reference estuaries).

Impacts of dredging-related resuspension were tested by the interaction terms Pe × PILvsRefs and Ti(Pe) × PILvsRefs (Tables 2 and 3). The first interaction indicates that there had been a consistent effect of the dredging (e.g. recruitment in Port Kembla was similar to that in the reference estuaries before dredging but not after dredging). The second interaction may indicate that an effect has occurred in only one of the two times during dredging, or alternatively, it may indicate that there was substantial variation between the PIL and references before dredging occurred. The first scenario would indicate an impact, while the second would not. Both interaction terms were further investigated (if significant) using pair-wise comparisons of the PIL and the references at each time (Underwood 1997).

Table 2.   Summary of the Beyond BACI asymmetrical anova to test for changes in sessile invertebrate recruitment in response to dredging contaminated sediments in the Inner Harbour of Port Kembla Thumbnail image of
Table 3.   Summary of the Beyond BACI asymmetrical anova to test for changes in sessile invertebrate recruitment in response to dredging contaminated sediments in the Outer Harbour of Port Kembla Thumbnail image of

All taxa that occurred in at least 20% of the samples (i.e. plates) were analysed. Taxa that occurred in fewer samples are difficult to analyse and interpret due to their patchy distributions. Inspection of box plots and residuals indicated all taxa could be analysed using untransformed data (Quinn & Keough 2002) with the exception of colonial ascidians. Data for these invertebrates were transformed to log10 (X + 0·01).

Summaries of the Beyond BACI analyses are presented showing the probability values for each of the terms in the analysis. Analyses of variance (anova) were carried out according to Underwood (1994). The initial anova were performed using systat 11 (Systat Software Inc., Chicago, Illinois, USA) and the asymmetrical Beyond BACI analyses were constructed in ms excel.

Turbidity data were analysed using a three factor anova comparing data before and during dredging in Port Kembla and Port Hacking. Metal concentrations were analysed using a two factor anova comparing data before and during dredging within Port Kembla.

Results

Sessile invertebrate recruitment

Inner Harbour

The recruitment of the major space-occupying species decreased dramatically in the Inner Harbour of Port Kembla with the onset of dredging, while no such decreases occurred in the reference estuaries (Fig. 2, Table 2). The extent of this change was most clearly demonstrated by the increase in bare space (the inverse of the total cover of invertebrates) in assemblages in the Inner Harbour which went from c. 5%‘before’ dredging to over 75%‘during’ dredging (Fig. 2, Table 2). This was an increase of bare space of over 155% relative to the reference estuaries.

Figure 2.

 Changes in sessile invertebrate recruitment in response to dredging-related resuspension contaminated sediments in the Inner Harbour of Port Kembla. Percentage covers of selected taxa are shown from the Inner Harbour [potentially impacted location (PIL) inline image mean only] and the reference estuaries (inline image mean + SE). See Statistical analysis section for further details.

The increase in bare space in the Inner Harbour was due largely to a marked change in recruitment of the dominant barnacle Amphibalanus variegatus (Darwin, 1854) with the onset of dredging activities (Fig. 2, Table 2). A. variegatus was relatively abundant before dredging commenced, covering on average between 40% and 50% of the recruitment plates (Fig. 2). It was less abundant in the reference estuaries, on average occupying less than 20% of the plates (Fig. 2), but this recruitment was relatively stable through time even during the dredging period. In the Inner Harbour, however, the recruitment of this barnacle was virtually eliminated during the dredging period (Fig. 2) suggesting strongly that an impact had occurred.

The serpulid polychaete, Hydroides elegans (Haswell, 1883), also showed a relative decrease in recruitment during dredging (Fig. 2, Table 2). This species was much more abundant in the Inner Harbour than at the reference estuaries before dredging commenced (Fig. 2, Table 2). When dredging began its cover increased in the reference estuaries but not in the Inner Harbour (Fig. 2, Table 2), and in the second round of sampling during dredging its cover decreased greatly in all of the estuaries (Fig. 2).

Several other invertebrates, such as the arborescent bryozoans, amphipods and colonial ascidians, showed patterns of recruitment suggesting that they may have also been affected by the dredging activities (Fig. 2). None of these patterns were, however, statistically significant (Table 2). The statistical power of Beyond BACI analyses for these taxa was low and, therefore, it was unlikely that an effect would be detected. In the case of the arborescent bryozoans and the amphipods, the low power was due to the great variation between the references estuaries, while in the case of the colonial ascidians it was due to the great variation through time at Port Kembla.

Several sessile invertebrate taxa [Diplosoma listerianum (Milne-Edwards, 1841), Herdmania grandis (Heller, 1878) and bivalves] showed significant fluctuations in recruitment between the Inner Harbour and the reference estuaries between sampling times (Table 2). Although such fluctuations could indicate pulse or delayed effects, this was not the case for any of these taxa as the differences only appear to exist prior to dredging commencing.

Outer Harbour

No changes in the recruitment of sessile invertebrates were observed in the Outer Harbour with the commencement of the dredging activities (Table 3). The major space occupier, A. variegatus, showed similar levels and patterns of recruitment to those at the reference estuaries, indicating no change in relation to the dredging activities (Fig. 3, Table 3). The amount of bare space did, however, suggest an overall change in invertebrate recruitment as there appeared to be a relative increase in bare space in the Outer Harbour assemblages compared with that in the reference estuaries (Fig. 3). This was, however, not statistically significant (Table 3).

Figure 3.

 Changes in sessile invertebrate recruitment in response to dredging-related resuspension contaminated sediments in the Outer Harbour of Port Kembla. Percentage covers of selected taxa are shown from the Outer Harbour [potentially impacted location (PIL) inline image mean only] and the external reference estuaries (inline image mean + SE). See Statistical analysis section for further details.

Turbidity

Turbidity in the Inner and Outer Harbours increased with the onset of the dredging operations (Fig. 4). Turbidity was, however, highly variable in Port Kembla and Port Hacking (Fig. 4) making it difficult to detect an interaction between estuary and period. Nonetheless, there was a significant increase in turbidity in both estuaries from before to during dredging (Period: F1,8 = 7·0, < 0·01) and this increase was greater in Port Kembla than in Port Hacking (Fig. 4).

Figure 4.

 Turbidity in the Inner and Outer Harbours of Port Kembla and reference estuaries. Mean (+SE) turbidity levels are shown for Port Kembla (inline image) and external reference estuaries (inline image).

In the Inner Harbour, there was a general increase in turbidity from before to during dredging of 2·3–6·9 ntu (Fig. 4). This increase appeared to bring the turbidity levels in the Inner Harbour into line with that in the upstream sections of Botany Bay (Fig. 4). In the Outer Harbour, turbidity was only slightly higher during dredging (0·8–2·0 ntu), except for a pulse of extremely turbid water (Fig. 4). During this pulse, turbidity measurements of 96 and 106 ntu were recorded within 20 m of the deposition area.

Aqueous metal concentrations

Manganese was the only metal which showed a clear increase in dissolved concentration during dredging (F1,38 = 7·82, < 0·01). This increase occurred in both the Inner and Outer Harbours (Fig. 5), although the relative change was greater in the Inner Harbour. Dissolved zinc also appeared to increase in concentration with dredging (Fig. 5). Zinc, however, was not detected in either the Inner or Outer Harbours in the first round of sampling in the ‘during’ period, despite being detected in the ‘before’ dredging sampling rounds (Fig. 5). This is presumably an artefact of the sampling process and does not reflect the complete absence of zinc at that time. No change was detected between periods for aluminium, arsenic and selenium (the other metals that were detected) within Port Kembla. Total suspended solids (TSS), measured in conjunction with the metals, also increased from the before to during dredging periods (F1,38 = 4·76, < 0·05). This increase was also substantially greater in the Inner than the Outer Harbour (Fig. 5). The increase in TSS paralleled the changes observed in turbidity within the harbour.

Figure 5.

 Aqueous metal concentrations and total suspended solids in the Inner and Outer Harbours of Port Kembla (mean + SE). Data: courtesy BlueScope Steel Pty Ltd (Port Kembla, NSW, Australia). Note: Zinc was not detected in Time 3.

Discussion

Dramatic decreases in the recruitment of sessile invertebrates occurred in the Inner Harbour of Port Kembla following the onset of dredging. No similar changes occurred at reference estuaries over the same period of time. The substantial decreases observed in Port Kembla were, therefore, consistent with an impact of the dredging operations (Underwood 1994; Downes et al. 2002). Furthermore, we argue that it was likely that this impact was driven by the resuspension of contaminated sediments and the possible release of pollutants into the water-column. No other activity relating to or concurrent with the dredging appears to offer an alternative, viable, explanation for the observed patterns of recruitment as no other large-scale disturbances were observed by ourselves, the PKPC, or government environmental regulatory authorities. For the remainder of this discussion, we will, therefore, consider the identified changes in recruitment in Port Kembla Harbour relative to the reference estuaries to be an impact of the resuspension of contaminated sediments.

Ecological disturbance

The large-scale resuspension of contaminated sediments in Port Kembla appeared to result in major ecological impacts. There were substantial increases in bare space and extensive decreases in the recruitment of the major space occupiers, particularly the barnacle A. variegatus and polychaete worm H. elegans. Impacts were, however, generally constrained to the Inner Harbour (where dredging occurred and most contaminated sediments were deposited) as there was little evidence of impacts in the Outer Harbour. Nevertheless, it appears that the impacts occurred over large spatial scales (100s of metres) within the Inner Harbour as samples were spaced representatively through the Inner Harbour and most were well away (>100 m) from the dredging activities. The impacts also occurred over relatively long temporal scales as the impacts on bare space and A. variegatus were found during both of the sampling times while dredging was occurring which covered a period of c. 4 months. The large spatial and temporal scales of these impacts on the recruitment of these sessile invertebrates suggest strongly that there had been a significant effect on the ecology of the estuary. The results of this study demonstrate that past pollution events can continue to cause current ecological impacts that can extend well beyond the area directly disturbed; in this case from the estuary-floor to a large section of the surrounding water-column.

The observed impacts on the dominant space-occupying invertebrate A. variegatus appeared to be the result of its recruitment being virtually extinguished from the Inner Harbour. There was almost no indication that A. variegatus had recruited during the dredging period in the Inner Harbour, with few live barnacles or even empty exoskeletons (which can remain attached to hard substrata long after the animal has died). This lack of recruitment indicates that the resuspension of contaminated sediments may have caused mortality to larvae in the water-column or early settlers (within a few days of settling), rather than killing older recruits or adults. Dredging-related resuspension may, therefore, have caused a press (i.e. virtually continuous) disturbance, as otherwise some larvae would be expected to have settled and metamorphosed between pulses of resuspension, leaving exoskeletons when they died. Consistent recruitment of A. variegatus has been observed here in previous years (Johnston 2006; Johnston & Clark 2007), supporting our interpretation that the decreased recruitment in Port Kembla was due to the dredging-related resuspension rather than seasonal variation.

Ecological impacts were most severe in the Inner Harbour, where dredging occurred and most of the heavily contaminated sediment was deposited. The Inner Harbour is smaller and experiences less tidal flushing than the Outer Harbour, and showed the largest increases in water-column turbidity and metal concentrations. The overall pattern of diminishing ecological impacts and variation in physicochemical conditions from the Inner to the Outer Harbour suggests that impacts were contained within Port Kembla and were unlikely to have passed outside of the estuary. This indicates that the management decision to retain the most contaminated sediments within the Inner Harbour was sensible.

Causes of the ecological impacts

The dredging impacts observed in the current study, we argue, were likely to be caused by the resuspension of the contaminated sediments of Port Kembla into the water-column, as it was the only major disturbance associated with dredging and deposition (Wilber & Clarke 2001). Potentially, these impacts could have been because of the release of a wide range of contaminants from the sediment, such as metals and PAHs (Table 1), as can occur regularly with the resuspension of contaminated sediments (Eggleton & Thomas 2004). Metals and PAHs may exist in a free ionic form or bound in complexes, although only the ionic form is usually considered bioavailable (Campbell 1995). Sulphides constitute the bulk of metals in anoxic sediments, and when disturbed they oxidize and release the metals as free ions (Simpson et al. 1998). Most fine sediments below a depth of 2–5 mm are anoxic (Simpson et al. 1998), which constituted virtually all of the sediments dredged in Port Kembla. These sediments also contained elevated levels of a range of metals and PAHs, many well above national environmental guidelines (Table 1, ANZECC 2000). Hence, the resuspension of these sediments may have released large quantities of metals and PAHs into the water-column (for examples of such events, see Simpson et al. 1998; Eggleton & Thomas 2004) and may have had toxic effects on the sessile invertebrates, as has been observed in small-scale manipulative experiments on phytoplankton and bacteria (Nayar et al. 2004).

Elevated concentrations of aqueous metals were, however, only observed for manganese and possibly zinc during dredging. It is possible that the limited detection of aqueous metals was a result of difficulties associated with snapshot sampling and pulse-type disturbances associated with the release of contaminants from dredging and deposition of sediment which may vary in frequency and intensity through time. Alternatively, metals and PAHs may have rapidly reformed complexes or may not have been released from the suspended sediment. This raises a significant issue of the importance of targeted chemical sampling compared with a routine sampling programme which may miss potentially ecologically important releases of contaminants (the latter possibly occurring during this study). Nevertheless, even bound contaminants would still have posed a threat when resuspended as they could be easily ingested by suspension-feeding invertebrates and while passing through the acidic gut of these animals they would have then become bioavailable, and potentially toxic (Gagnon & Fisher 1997; Ke & Wang 2002). It is also possible that both free ion uptake and ingestion may have occurred simultaneously (Gillis et al. 2006).

A bioaccumulation study (Hedge, Knott & Johnston 2009) conducted simultaneously with the current study found that oysters (Saccostrea glomerata, Gould, 1850) deployed in Port Kembla accumulated a range of metals in higher concentrations during dredging compared with before dredging, relative to oysters deployed in reference estuaries (i.e. Botany Bay and Port Hacking). This pattern was most evident in the Inner Harbour, where oysters accumulated copper and zinc at levels two and three times greater during dredging than before (Hedge, Knott & Johnston 2009). These findings clearly support the model that contaminants played a role in the ecological impacts observed in the current study, however, they provide no indication of whether these effects were via aqueous uptake or ingestion, or both mechanisms simultaneously. Future research assessing the main uptake pathways for contaminants from resuspension events should be a priority.

It is also possible that impacts on marine invertebrates were caused by increased suspended sediment loads, which can damage or block feeding and respiratory organs (Airoldi 2003; Lohrer, Hewitt, & Thrush 2006; Maldonado, Giraud, & Carmona 2008). The turbidity levels recorded in the current study were not, however, greatly elevated (i.e. TSS 6 mg L−1) in comparison with other estuaries locally (e.g. Botany Bay) or worldwide (Wilber & Clarke 2001). They were also much lower than that known to cause mortality in laboratory studies (e.g. >400 mg L−1 for oyster larvae and adults, Wilber & Clarke 2001). Hence, we argue that the suspended sediment loads were not increased to a biologically significant level and that it was the contaminants in, or released from, the sediments that were likely to be responsible for the impacts observed in the current study and not the result of a slight increase in the suspended sediment load.

Management of dredging-related resuspension and ecological assessment

Dredging operations in many parts of the world are carried out with great attention to potential environmental effects (Environment Australia 2002; USEPA 2006). There have been substantial technological and managerial advances in the minimization of dredging-related effects (Herbich 2000; Munns et al. 2002; Suedel et al. 2008), many of which are aimed at reducing or containing turbidity (Wilber & Clarke 2001; Eggleton & Thomas 2004; Erftemeijer & Lewis 2006). The current study suggests that despite good management and the use of sophisticated technology (e.g. low energy diffusers for sediment deposition), dredging in Port Kembla led to significant ecological impacts. This indicates that the large-scale resuspension of contaminated sediments is, therefore, a real threat and needs to continue to be managed carefully.

Great care and attention should be taken with the containment of resuspension events (i.e. via the use of silt curtains) and additional engineering solutions may be required to further curtail the resuspension events caused by dredging operations in areas where there is significant concern about potential ecological impacts. Dredging techniques that cause less disturbance to the sediments (specifically large-scale mixing of contaminated anoxic sediments with seawater) may be required. These may include sweephead dredging (Eggleton & Thomas 2004), or dredgers that cut whole blocks of intact sediments. Nevertheless, understanding ecological impacts is first required to make appropriate, cost-effective and logical decisions about the management of such anthropogenic disturbances, and the current study has contributed substantially to the development of this knowledge.

Our field-based assessment using a Beyond BACI design was relatively simple and easy to perform. The simultaneous sampling of two external reference estuaries allowed impacts of the large-scale resuspension of contaminated sediments to be distinguished from background and seasonal variation in invertebrate recruitment (Underwood 1996; Downes et al. 2002). Importantly, sessile invertebrate recruitment incorporated both the sensitive settlement and early adult phases of the life-histories of these organisms, which are key components of estuarine ecology. It also enabled a cost-effective assessment of impacts of resuspended contaminated sediments on a range of species in several life-history stages, which would have been difficult to replicate had standard ecotoxicological tests been used.

The current study of recruitment was also considerably more efficient than other field-techniques such as in situ subtidal surveys. Deployment of recruitment plates avoids SCUBA diving at potentially polluted sites, and the recruitment assemblages are relatively quick to census and incorporate a broad cross-section of phyletic diversity. Recruitment studies have a long history in ecology (Pomerat & Reiner 1942; Lewis 1976; Young 1990), and appear extremely useful for the assessment of pollution events in marine environments. Overall, the current study demonstrates the usefulness of such techniques and the importance of BACI-type experimental designs to assess whether threats predicted in laboratory studies have real-world impacts. It provides the clearest evidence to date of large-scale ecological impacts of resuspended contaminated sediment.

Acknowledgements

We thank L. Hedge, K. Stuart, M. Black, K. Dafforn, J. Lawes, D. Cruz, D. Bolton and H. Durrant for assistance with fieldwork and sampling of sessile invertebrate assemblages; the Port Kembla Ports Corporation financial support for this project and for facilitating access to all locations within the Port Kembla throughout the dredging operations; and the helpful comments from G. Clark, J. Gill and two anonymous reviewers. This study was partially funded by an Australian Research Council Linkage grant awarded to E. Johnston, N. Knott and T. Brown.

Ancillary