Copper (Cu) is a widely used metal around the world. The global annual consumption of Cu in 2005 was more than 15,000 metric tons, second only behind aluminum (Rogich and Matos 2008). A fraction of this Cu enters the aquatic environment, where it readily interacts with other chemicals that are dissolved or associated with particulate matter in the water column (Phillips et al. 2004). As the Cu-containing particulate matter settles to the sediment, complexes are formed between Cu and various geochemical ligands of different binding capacities. These Cu-ligand complexes collectively determine the speciation of Cu between a state that is bound to the bulk phase of sediment and a dissolved state in the pore water (PW) or overlying water (OW). Dissolved Cu is more readily bioavailable, as Cu2+ may directly partition to gill tissue of aquatic organisms, allowing for its entry into the body. Speciation between the ligand-bound and dissolved states is dependent upon sediment redox potential, pH, concentration of acid volatile sulfides (AVS), total organic carbon (TOC), carbonates, and oxyhydroxides of iron (Fe) and manganese (Mn) (Burton and Kapo 2007).

Given the economic importance of Cu and its presence in marine sediments, it is important that its bioavailability and toxicity to marine life be determined under conditions that simulate natural conditions at the sediment/water interface. Determinations of acute and chronic effects upon benthic animal species of different sensitivities and ecological niches, under normal reducing redox conditions within the sediment, are important for the development of meaningful sediment quality guidelines (SQGs) for the protection of marine life. Laboratory studies that have been conducted under conditions within the sediment that are more oxidizing than what normally occurs in situ, or that have not allowed sufficient time for equilibrium of Cu concentrations to occur amongst the bulk sediment, PW and OW phases, may have overestimated toxicity (Simpson et al. 2012) due to the release of ligand-bound Cu2+ into the PW or OW (Hutchins et al. 2008, 2009).

The present study was conducted to determine the concentrations of Cu in sediment that adversely affect the marine amphipod, Leptocheirus plumulosus, in a chronic laboratory exposure simulating natural conditions. It was performed as a companion study to a chronic exposure of Cu to the marine polychaete, Neanthes arenaceodentata (Ward et al. 2015).

Materials and Methods

The test chemical, CuCl2·H2O (Acros Organics, Fairlawn NJ, USA; 100.5 % purity), was used to prepare all solutions for the toxicity test, matrix spiking tests, and analytical standard solutions. Sediment from the Pacific Ocean (Sequim Bay, WA, USA) was collected on July 11, 2011, shipped in coolers to ABC Laboratories (Columbia, MO, USA), and stored at ~4°C until test initiation on October 17, 2011. It was not sieved prior to use. Sediment texture measurements were determined by AGVISE Laboratories (Northwood, ND, USA) using a hydrometer method, and yielding a composition of 42 % sand, 40 % silt, and 18 % clay. Several parameters were determined by Inovatia Laboratories (Fayette, MO, USA). These included a measure of carbonate content at 0.6 % (EPA method 310.1, USEPA 1983); measurements of Fe, Al and Mn oxyhydroxides at 181, 54.7 and 1.99 mg/kg dry sediment, respectively, following the sequential extraction procedure of Shannon and White (1991) and metal analysis by EPA Method 200.8 (USEPA 1994); TOC at 0.758 % according to a high-temperature combustion method (APHA 2006); and total metal concentrations as determined by inductively coupled plasma-mass spectrometry (ICP-MS) in a single sample of sediment using EPA Method 200.8 (Table 1). Acid-volatile sulfide measurements (Table 2) were performed by Wilfred Laurier University (Waterloo, ON, Canada) using the weak acid extraction method of Allen et al. (1991). Pore water salinity ranged from 19.6 to 21.2 ‰. Water used for acclimation of test organisms and in the toxicity test was GP2 artificial seawater (Arnold et al. 2007), adjusted to a salinity of 20 ± 2 ‰.

Table 1 Metal concentrations in control sediment sample (mg/kg dry sediment)
Full size table
Table 2 Mean concentrations of copper in sediment, pore water (PW) and overlying water (OW); sediment acid-volatile sulfide concentrations [AVS], sediment [total Cu]-[AVS] differences; redox potential in sediment and OW; and survival, growth and reproduction of Leptocheirus plumulosus after 28 days of exposure
Full size table

Nominal concentrations of Cu in sediment for the toxicity test were 40, 80, 160, 320 and 420 mg Cu/kg dry sediment, plus control (no Cu added). The spiking procedure and test conditions used to simulate natural equilibrated sediment with a reducing environment and slightly alkaline pH were previously described (Gaertner 2012; Ward et al. 2015). Briefly, sediment was first spiked at a high concentration (superspike of 63,400 mg/kg dry sediment) under a nitrogen (N2) atmosphere, and allowed to equilibrate for 112 days. At 12 days prior to test initiation, the super-spiked sediment was diluted with unspiked sediment under a N2 atmosphere, and allowed to equilibrate at room temperature. Three days prior to the start of the test, overlying water (OW; ~725 mL of artificial seawater) was added under N2 to each exposure jar containing ~175 mL (263 g) of sediment, and gentle aeration of the OW was initiated. The jars were randomly placed into a temperature-controlled water bath at 25 ± 2°C under a 16:8 h light:dark photoperiod with 30-min transition periods at dawn and dusk.

Neonates of L. plumulosus (0.25–0.5 mm in length) were received from Chesapeake Cultures (Hayes, VA, USA) 2 days prior to testing for acclimation to test conditions and dilution water, during which time they were fed a flake food suspension. The toxicity test followed a protocol based on the USEPA-USACE (2001) method, with modifications in food (algae added) and frequency of water renewal (once weekly). Briefly, 20 healthy animals were randomly placed into each of five replicate 1-L glass jars containing Cu-spiked sediment at five exposure levels, plus control, for a 28-day static renewal exposure (overlying water was renewed weekly). The replicates received approximately equal volumes of suspended solids (22.6 g/L of Tetramin® and 6.5 g/L algae) three times weekly. After 28 days of exposure, live adult and juvenile amphipods were collected, enumerated and weighed. Three separate sets of 20 animals were dried overnight at ~70°C, and weighed for initial weight. All surviving adult amphipods at the end of the 28-day exposure were similarly dried and weighed for growth determination. Reproductive output was determined as the number of offspring per treatment replicate and per surviving adult.

Measurements of temperature, dissolved oxygen (DO) concentration, pH, and salinity of the overlying water were measured in at least one test chamber per treatment at least three times per week according to procedures described in a companion study (Ward et al. 2015). The water bath temperature, ammonia concentrations and light intensity were also measured during the test as previously described (Ward et al. 2015).

Copper concentrations were measured in sediment, OW and PW by ICP-MS using an Agilent Model 7500A instrument (Agilent Technologies, Santa Clara, CA, USA). Samples of sediment, OW and PW were collected at the beginning and end of the test from duplicate test vessels that did not contain organisms. For OW and PW analysis, 10-mL samples were collected from the control and test treatments. Overlying water samples were collected by transfer of OW from the test chambers into 15-mL culture tubes. After complete removal of OW, the sediment was centrifuged for approximately 20 min at 19,000 RPM, and the supernatant was transferred to Falcon tubes for PW analysis. A 0.100-mL aliquot of each sample was diluted to 10-mL with 2 % HNO3 to provide final sample concentrations within the concentration range of the analytical standards. Following PW removal, sediment moisture was analyzed, and approximately 5 g dry weight of centrifuged sediment was collected from the control and treatments and placed in reaction vessels. A 2.5-mL volume of extraction solvent (nitric acid) was added to each vessel and the samples were brought to a volume of 5-mL with reagent water. The samples were mixed, and heated to 95 ± 5°C for approximately 1 h. Upon cooling to room temperature, the volume was adjusted to 5 mL with reagent water. Aliquots of the supernatant were diluted with reagent water to provide final sample concentrations within the range of the analytical standards. All samples were directly analyzed using ICP/MS. Recovery ranges for Cu- fortified saltwater were 93 %–99 % at 0.600 mg/L, 76 %–88 % at 1.20 mg/L, and 108 %–111 % at 131 mg/L. For Cu-fortified sediment at 48.0 mg Cu/kg sediment, the recovery range was 67 %–88 %. Measured concentrations were not adjusted for recoveries.

Shapiro–Wilk’s and Levene’s tests were used to determine normality and homogeneity of variance, respectively. Survival, growth and reproduction data were normally distributed and variances were homogeneous, and a one-way analysis of variance (ANOVA) and parametric Dunnett’s test were used to calculate LOEC and NOEC values (p < 0.05). Statistical analyses were performed using SAS software (SAS Institute, Inc., Cary, NC, USA). Concentrations that inhibited endpoints by 50 % (i.e., the IC50) were estimated by linear interpolation software (USEPA 1993).

Results and Discussion

Copper concentrations in the 5 treatment groups ranged from 44.4 to 605 mg Cu/kg dry sediment (Table 2). The background control sediment contained a mean of 40.5 ± 4.58 mg Cu/kg sediment, similar to the pre-industrial concentration of 34.3 mg/kg for Puget Sound sediment (Brandenberger et al. 2008). The other element for which a comparison was possible was Pb, with concentrations of 10.4 and 10.7 mg/kg in current and pre-industrial sediment, respectively. Copper concentrations in the sediment were consistently maintained throughout the study. A comparison of mean measured Cu concentrations of sediment in the treatment groups on days 0 and 28 showed that day 28 concentrations were 96 %, 86 %, 103 %, 75 %, and 94 % of day 0 concentrations for the respective treatment groups in ascending order. Negative redox potential readings in the sediment indicated that a reducing environment was maintained over the exposure period. Copper concentrations in PW were 20.0, 25.0 and 60.0 µg/L in the three highest exposures. Mean concentrations in PW were ≤5.5 µg Cu/L in the control and two lower treatment groups. Copper concentrations in the OW were more variable overall, but were ≤16.6 µg Cu/L for all exposures except for the highest. Positive and consistent redox readings from 59.4 to 72.3 mV were obtained in the OW.

Un-ionized ammonia concentration ranges in the OW during the test for controls and all treatment groups were 0.368-0.569 and <0.0001–0.002 mg/L on days 0 and 28, respectively. Ranges for other OW water quality parameters over the 28-day exposure were: temperature, 24.4–25.0°C; pH, 7.1–8.5; salinity, 19.6–21.2 ‰; and DO, 4.8–8.3 mg/L (≥66.7 % of saturation at 25.0°C and 20 ‰ salinity). Un-ionized ammonia concentration ranges in the PW for all groups were 0.384–0.729 and 0.032–0.083 mg/L on days 0 and 28, respectively. The pH in PW over the 28-day exposure ranged from 7.56 to 7.89, and salinity ranged from 20.5 to 23.7 ‰. Light intensity at the surface of the overlying water was 925–1180 lux. All water quality parameters were within a normal safe range, and did not adversely impact the study results.

There was no evidence of a dose–response relationship for the survival endpoint, and no significant effects upon survival were evident at any Cu exposure level (Table 2; Fig. 1). In a 42-day chronic exposure of the amphipod, Melita plumulosa, to Cu-spiked sediment, the fertility endpoint was the most sensitive, followed by gravidity, length and lastly survival (Gale et al. 2006). The relative insensitivity of the survival endpoint in that study is in agreement with our observations. Based upon results in the present study, effects upon survival would occur at concentrations >605 mg Cu/kg dry sediment. In a companion study with the polychaete N. arenaceodentata (Ward et al. 2015), survival was not affected at 505 mg Cu/kg sediment, but was significantly reduced at 1190 mg Cu/kg sediment. The mean control survival of 75 % did not meet the test protocol’s acceptability criterion of 80 % for this endpoint (USEPA-USACE 2001). One of the 5 control replicates had an inexplicably low survival of 45 %, lowering the overall mean survival rate. Mean survival for the other 4 control replicates was 82.5 %. Difficulties in meeting this criterion in tests with this species have been experienced by other laboratories. It was recently suggested that elevated nitrite levels in the test systems may be responsible for control mortalities (Bradley et al. 2014). They found that improved control survival may be achieved by the addition of nitrifying bacteria into the test system. Another means of reducing potentially toxic nitrite levels would be to allow additional time for the normal nitrogen cycle to dissipate ammonia, nitrite and nitrate from the test system before adding animals. Ammonia concentrations in PW are typically the highest at the start of tests (Bradley et al. 2014; Ward et al. 2015; this study). We allowed 3 days between the addition of OW and animals into the test system. In spite of the failure to meet the 80 % control survival criterion, the present study contains valuable information on the effects of Cu-spiked sediment upon L. plumulosus in an environment simulating natural conditions. It is important that protocols be revised as new information becomes available to facilitate laboratory compliance with performance acceptability criteria in these lengthy and expensive tests.

Fig. 1
figure 1

Survival, growth and reproduction of Leptocheirus plumulosus following exposure to Cu-spiked sediment for 28 days. Error bars = mean SE; asterisk indicates a significant difference (p < 0.05) relative to background control

Full size image

Adverse biological effects were first observed in the growth endpoint, followed by the reproduction endpoint, when evaluated by hypothesis testing. Growth was reduced (p < 0.05) at the treatment level of 418 mg Cu/kg dry sediment, but not at the next lower level of 199 mg Cu/kg sediment (Table 2; Fig. 1). The highest exposure of 605 mg Cu/kg sediment caused significant reductions in both growth and reproduction. The NOEC–LOEC range, based upon growth as the most sensitive endpoint, was 199–418 mg Cu/kg dry sediment. A significant reduction in growth was observed at a concentration of 505 mg Cu/kg dry sediment with N. arenaceodentata (Ward et al. 2015). Growth was also the most sensitive endpoint in that study, and the NOEC-LOEC range was 230–505 mg Cu/kg sediment.

Reproduction expressed on the basis of total offspring per replicate was significantly reduced (p ≤ 0.05) at the highest exposure level of 605 mg Cu/kg dry sediment (Table 2; Fig. 1), when analyzed with a hypothesis test. The reduced levels of reproduction at the next two lower levels were similar (47 % and 54 %), but were not statistically significant. Analysis of total offspring per replicate by linear interpolation yielded an IC50 estimate of 187 mg Cu/kg dry sediment (CL = 153–473 mg/kg). When reproduction was expressed on the basis of offspring per surviving adult, none of the treatments resulted in a significant reduction from the control, and the IC50 estimate was 492 mg Cu/kg dry sediment. In a recent study, Eickhoff et al. (2014) reported that the reproduction endpoint with L. plumulosus was more variable than the growth endpoint, and that this variability decreased with increased growth of test animals. They proposed a mean dry weight threshold of 1.3–1.5 mg per control animal to reduce variability around the reproduction endpoint. A high degree of variability in the reproduction endpoint was also reported in the USEPA-USACE protocol (2001). The final mean dry weight of our control animals was 0.977 mg, a weight that is common in 28 day exposures with this species, but which is considerably less than the weight recommended by Eickhoff and colleagues.

Reproductive endpoints in amphipod toxicity tests are useful in assessing risk to field populations, since reproductive capacity determines population growth rates and ultimate population survival in the field (Gale et al. 2006; Hyne 2011). Reproductive endpoint information from tests with L. plumulosus was used to develop interpretive guidance for a population model in the Baltimore Harbor, USA (McGee and Spencer 2001). Based upon results of L. plumulosus exposure to Baltimore Harbor sediment contaminated with Cu plus other metals and organic contaminants, Manyin and Rowe (2006), suggested that energy was being diverted from productivity to maintenance pathways, resulting in slower growth and reduced fecundity. The EC50 for fertility was 290–330 mg Cu/kg dry sediment in three chronic tests with the amphipod, M. plumulosa (Gale et al. 2006). Considering the importance of the reproduction endpoint in the extrapolation of toxicity test results to possible effects in the field at the population level, an evaluation and possible revision of the test protocol is recommended in an attempt to increase amphipod growth and decrease variability associated with the reproduction endpoint.

From a large database of sediment chemistry and toxicity data (decreased survival) for Cu and the North American marine amphipods Ampelisca abdita and Rhepoxynius abronius, Field et al. (2002) predicted threshold toxicity concentrations at the 20 %, 50 % and 80 % probability levels to be 32, 94 and 280 mg Cu/kg dry sediment, respectively. While we did not observe a significant decrease in survival, the LOEC for growth (418 mg Cu/kg dry sediment) of L. plumulosus exceeded the predicted 80 % effect threshold level. It also exceeded the AET (apparent effects threshold) sediment quality guideline (SQG) benchmark of 390 mg Cu/kg dry sediment, which is the highest of the various empirical benchmarks listed by Buchman (2008). However, the IC50 estimate for total reproduction per replicate (i.e., 187 mg/kg dry sediment) fell between the 50 % and 80 % probability levels for the predicted toxicity threshold. Comparison of the LOEC of this study with mechanistic SQG benchmarks which normalize for AVS and TOC (USEPA 2005; Burgess et al. 2013), indicated that toxicity would be “unlikely” for the lowest sediment exposure (44.4 mg Cu/kg dry sediment) and “uncertain” in the others. Another mechanistic approach normalizes for OC in the <63 µm silt plus clay particle size fraction (Simpson et al. 2011). This approach would predict a safe concentration of 26.5 mg Cu/kg dry sediment for our test sediment, if all of the OC was present in the <63 µm fraction, and <26.5 mg/kg if some of the OC was associated with larger particles. Prediction of a safe level using either of these approaches suggests a high degree of protection for L. plumulosus. In fact, the safe level for sediment of this study would appear to correspond to Cu levels that were characteristic of Puget Sound sediment from pre-industrial time.

It should be noted that the molar concentration of Cu in Table 2 (i.e., [Cu]) is based on a measure of total Cu, and not Cu extracted simultaneously with AVS. Such SEM [Cu] would be lower, perhaps by approximately 50 %. In a comprehensive study in two rivers of the Pacific Northwest, Patton and Crecelius (2001) showed from a linear regression analysis that [SEM Cu] was approximately 56 % of [total Cu].

Pore water concentrations of Cu at the two lowest exposures were similar to the control (Table 2), indicating that Cu was in equilibrium between PW and sediment. Concentrations of Cu in PW at the two highest exposures were 25 and 60 µg/L, where reductions in growth and reproduction occurred (Table 2; Fig. 1). At the intermediate sediment concentration of 199 mg Cu/kg sediment, the PW concentration was 20 µg Cu/L. Reproduction was reduced, but not significantly here, possibly indicating an incipient sublethal effect level. In a companion study with N. arenaceodentata (Ward et al. 2015), Cu concentrations in PW of 38.7 and 65.8 µg/L reduced growth and survival, respectively; and effects were not observed ≤22.4 µg Cu/L. Strom et al. (2011) observed complete mortality in a sediment toxicity test with the amphipod Melita plumulosa in a 10-d exposure with a PW concentration of 41 µg Cu/L. The PW concentrations of Cu where adverse effects occurred in the present study were well above the current US saltwater benchmark concentration of 3.12 µg/L (USEPA 2005; Burgess et al. 2013). This PW benchmark concentration would appear to be highly protective of L. plumulosus over its entire life cycle.