Two forms of ammonium exist in the aquatic environment: un-ionized ammonia NH3 and ionized ammonium NH +4 . The toxicity to many aquatic organisms is primarily attributed to the NH3 (un-ionized) species, with the NH +4 ion (ionized) species being relatively less toxic (Williams et al. 1986; Kohn et al. 1994; Borgmann 1994). Factors such as pH (Erickson 1985), temperature (Erickson 1985), and oxygen saturation (Wasjbrot et al. 1990), which influence the degree of ionization, are therefore of critical importance in determining total ammonium toxicity. However, some authors state that the toxicity of total ammonium at low pH (< 8.2) levels is due not only to un-ionized NH3, but also to ionized ammonium NH +4 (Armstrong et al. 1978; Landau and Sanchez 1991; Schubauer-Berigan et al. 1995).

The marine amphipod Corophium volutator is used as a bioassay species to assess the toxicity of marine sediments and dredged materials in several European countries, including the Netherlands (Stronkhorst et al. 2003), the United Kingdom (Bat and Raffaelli 1998), Germany (Peters et al. 2002), and Denmark (Pedersen et al. 1998). High levels of (total) ammonium in these toxicity tests can potentially confound bioassay results (Moore et al. 1997). In these circumstances, the contribution of priority pollutants to toxicity cannot be accurately quantified unless the potential contribution of total ammonium to toxicity is first ruled out (Borgmann and Borgmann 1997). In a Corophium bioassay on sediments originating from a harbor or another field location, temperature and salinity are kept constant and the overlying water is saturated with oxygen. The pH level depends on the sediment, however, and is less easy to control in the bioassay. This makes it very important that quality criteria be derived for total ammonium levels at several pH values in sediment bioassays such as the Corophium bioassay. These quality criteria should preferably address concentrations in the overlying water, because it is difficult to perform routine measurements in pore water during the bioassay without disturbing both sediment and bioassay.

Armstong et al. (1978) have hypothesized that, at pH levels less than 8.3, both NH3 and NH +4 play a role in total ammonium toxicity. The goal of the present research was to determine whether additional toxicity of the ionized form NH +4 can be detected at pH values greater than 8, to aid setting of water quality criteria for this confounding factor in Corophium sediment bioassays. Ammonium tests mostly take the form of 24-hour to 96-hour water-only tests (e.g., Arthur et al. 1987; Kohn et al. 1994; Richardson 1997). Sometimes water-only tests using an ammonium concentration range are performed over a longer period: one week (Borgmann and Borgmann 1997), 2 weeks (Williams et al. 1986) or even 10 weeks (Borgmann 1994). However, the absence of sediments might easily constitute an extra stress factor, biasing the result, especially for sediment-inhabiting organisms like Corophium. Therefore, next to water-only experiments, organisms were exposed in a pH-controlled sediment-water test system.

Materials and Methods

Corophium volutator

The amphipods used in the present study were collected at Oesterput, a relatively unpolluted site located in the Eastern Scheldt (the Netherlands). The animals were collected by sieving (500-μm mesh) sediment from the intertidal zone. The remaining fraction, containing the animals, was transported to the laboratory, where the Corophium were acclimatized to the standard temperature of 15°C for at least 2 days. They were not fed. Only animals not able to pass through a 500-μm mesh were used in the experiments. All experiments were performed within 10 days of collecting the animals in the field.

pH-Controlling

In all experiments, the pH was controlled using the carbonate buffer system. In this system, carbon dioxide is added to or extracted from the air used for aeration (to raise or lower the pH, respectively). This form of pH manipulation in marine toxicity tests is described by Mount and Mount (1992). The method has been adjusted for the testing of larger systems by using continuous aeration with a (carbon dioxide) manipulated mixture of air and carbon dioxide, using flow meters for regulation. Glass beakers containing a water or sediment–water system were placed in the experimental chamber, in which the desired pH level was reached after 24 hours. In all experiments, pH was controlled twice a day and adjusted if it differed from the desired pH by more than 0.1.

pH-Controlled Water-Only Tests

Toxicity tests with ammonium chloride (NH4Cl; Acros Organics, Geel, Belgium) dissolved in sea water (32 psu) were performed to evaluate the response in water-only tests. Corophium were exposed for 72 hours to a series of concentrations, using three replicates of each concentration. Depending on expected sensitivity, nominal ammonium concentrations varied between 0 mg/l and 100 mg/l at low pH and between 0 mg/l and 45 mg/l at high pH. Ammonium chloride was diluted in seawater from the Eastern Scheldt filtered over a sand-bed filter to remove particles larger than 10 μm. Experiments were performed in 1-litre glass beakers. Twenty randomly selected Corophium were added to each beaker at the beginning of the test. Test solutions were aerated continuously during the experiment. The following parameters were measured: salinity and oxygen saturation at the beginning and end of each experiment, and temperature, pH, and total ammonium concentration on a daily basis. Total ammonium was measured using the indophenol blue spectrophotometric method (Dr. Lange GmbH, Düsseldorf, Germany). After 72 hours (70–76 hours), the number of surviving Corophium in each beaker was ascertained.

pH-controlled water-only tests were performed during the winter of 2000–2001. Table 1 gives an overview of the experiments performed and their dates. A water-only test without controlled pH was performed four times in winter and four times in summer to indicate any variation in the fitness of the batches of Corophium used.

Table 1 Dates of sediment and water-only tests
Full size table

pH-Controlled Sediment Tests

Sediment toxicity tests (10-day) with ammonium chloride (NH4Cl; Acros Organics, Geel, Belgium) dissolved in sea water (32 psu) as overlying water were performed to evaluate the response in sediment bioassays. The sediment used in the experiments was collected at Oesterput, the location where the Corophium were collected. This sediment is known for its low natural ammonium content. Sediments were sieved over 500 μm in the field to remove organisms present and transported to the laboratory. Experiments were performed in 1-litre glass beakers, with three replicates per concentration. Depending on expected sensitivity, nominal ammonium concentrations varied between 0 mg/l and 320 mg/l at low pH and between 0 mg/l and 100 mg/l at high pH. Each beaker was filled with 200 ml sieved sediment, and 600 ml sea water with the desired total ammonium concentration. To obtain the desired concentrations, ammonium chloride was diluted in seawater from the Eastern Scheldt filtered over a sand-bed filter to remove particles larger than 10 μm. After 24 hours, 20 randomly selected Corophium were added to each beaker, and exposed to the concentration series for 10 days. The overlying water was aerated continuously during the experiment. The following parameters were measured: salinity and oxygen saturation at the beginning and end of each experiment, and temperature, pH, and total ammonium (for method see Water-Only Tests) on a daily basis. Total ammonium concentrations tend to decrease during the experiment; when measured total ammonium were more than 10% lower compared to the ammonium concentration on day 0, the ammonium concentration was adjusted. This was done by calculating which volume from a stock of 10 g NH4CL/l was needed to adjust the ammonium concentration in the experimental water volume of 600 ml, and adding this volume. After 10 days, the number of surviving Corophium in each beaker was ascertained. pH-controlled sediment tests were performed during the winter of 2000–2001. Because Corophium is known for its seasonal variation in response to cadmium (Kater et al. 2000), a second series of pH-controlled sediment tests was performed in the summer of 2001. Table 1 gives an overview of the experiments performed and their dates.

Calculations and Statistics

For each water-only and sediment test, the LC50 was computed, after logarithmic transformation of the concentration data, using the trimmed Spearman Karber method (Hamilton et al. 1977) and TOXCALC (Tidepool 1995).

Possible differences in responses during the winter and summer were tested using a Kruskal-Wallis test.

Assuming only NH3 and no NH +4 effect at a pH level greater than 8.3, all total ammonium concentrations of test performed at a pH > 8.3 were recalculated to NH3 concentrations, using the following equation (Whitfield 1974):

$$ \%\,{\rm NH}_3\,=\,100\,*\,(1\,+\,{\rm antilog}\,{\rm (pKa}\,(=\,dissociation\,constant)\,-\,{\rm pH)\,-\,1)} $$

All single NH3 concentrations from all experiments and their effect data were combined. One dose–effect relationship was calculated using the trimmed Spearman-Karber method. The LC50 value of this overall dose–effect relationship was calculated. This LC50 value was used to calculate an expected total ammonium toxicity (based on single NH3 toxicity), and confidence interval, at all pH levels between 8.0 and 9.0. The LC50 based on the individual sediment experiments was compared to this expected total ammonium toxicity.

Results and Discussion

General Performance of the Tests

In all experiments performed survival in control sediments (i.e., sediments without ammonium addition) was more than 90%, even at a pH of 9.0. Corophium was found not to be sensitive to high pH levels. In control experiments, total ammonium concentrations were measured on a daily basis. The average total ammonium concentration in 10-day sediment experiments varied between 1.7 and 5.3 mg/l. The highest total ammonium concentration measured in a control was 7.5 mg/l (day 10 at pH = 8.1 sediment experiment). The average temperature in sediment experiments varied between 13.8 and 15.9°C, whereas the average oxygen saturation in sediment experiments varied between 86% and 99%, and average salinity in sediment experiments varied between 28.7 and 33.8 psu. The pH levels in the experiments were kept constant and never varied more than 0.1 unit. Total ammonium concentrations were measured on a daily basis. Total ammonium concentrations tended to decrease during the experiment, but daily adjustment of the concentration kept this variation less than 10%. All calculations were performed with measured total ammonium values.

pH-Dependent Toxicity in Water-Only Experiments

From a total of eight 3-day water-only pH-controlled experiments, it was possible to calculate LC50 values and 95% confidence intervals using the trimmed Spearman-Karber method. Figure 1 gives the results of these pH-controlled water-only experiments. After logarithmic transformation of LC50 values, they showed a significant decrease with increasing pH (p < 0.001). Other studies have shown that sensitivity to ammonium is affected by life-stage: the younger or smaller the individuals, the more sensitive the organisms are to ammonium (Young-Lai et al. 1991; Lin et al. 1993; Moore et al. 1997). Noack et al. (2003) showed that Corophium with different body lengths (4–8 mm) showed no significant differences in sensitivity to ammonium. In this study, possible age effects were reduced by selecting organisms on inability to pass through a 500-μm sieve, and were therefore assumed not to play an important role.

Figure 1
figure 1

pH-dependent response of Corophium volutator to total ammonium in eight pH-controlled water-only tests

Full size image

pH-Dependent Toxicity in Sediment Experiments

The LC50 and 95% confidence interval were calculated from a total of six 10-day sediment pH-controlled experiments in the winter period, and six in the summer period, using the trimmed Spearman-Karber method. A Kruskal-Wallis test did not show a significant difference in response between winter and summer (p = 0.631; n = 12). Figure 2 gives the results of these pH-controlled sediment experiments. After logarithmic transformation of LC50 values, they showed a significant decrease with increasing pH (p < 0.001). LC50 values varied, depending on pH, between 30 and 80 mg/l, based on overlying water concentrations. At a pH of around 6.8, which is lower than the lowest pH of 8.1 in this study, the LC50 in a 10-day spiked sediment test was 43.3 mg/l for Lumbriculus variegatus, 87.0 mg/l for Chironomus tentans, and 9.7 mg/l for Hyalella azteca (Whiteman et al. 1996). Although the spiking method differs from the methods used in this study, Corophium seems less sensitive than the abovementioned species.

Figure 2
figure 2

pH-dependent response of Corophium volutator to total ammonium in 12 pH-controlled sediment tests

Full size image

Toxicity in Sediment Tests and in Water-Only Tests

When the LC50 of 10-day sediment tests was comparable to 3-day water-only test responses, the quotient between LC50 in sediment and LC50 in water was 1.03 for pH = 8.3 and pH = 8.5. At a pH of 8.7, this quotient increased to 1.69. Parallel 10-day water-only and 10-day sediment tests showed a higher LC50 for water-only tests compared to sediment tests with the oligochaete Lumbriculus variegates (LC50 sediment/ LC50 water = 7.3; Whiteman et al. 1996) and the midge Chironomus tentans (LC50 sediment/ LC50 water = 6.1; Whiteman et al. 1996). On the other hand, however, the amphipod Hyalella azteca showed a higher LC50 in the sediment test (LC50 sediment/LC50 water = 0.9; Whiteman et al. 1996). It seems that LC50 values in sediment tests are higher on the whole than in water-only tests. In the case of Corophium, this can be explained by the stress experienced by a benthic organism in a water-only test.

NH +4 Versus NH3 Toxicity

All sediment experiments with a pH greater than 8.3 were used to estimate an overall LC50 for NH3. The NH3 concentration was calculated from each individual total ammonium concentration in sediment experiments with a pH greater than 8.3. The calculated NH3 concentrations ranged from 0.5 to 9.6 mg/l. A dose–effect relationship was derived from this set of data. The LC50 of NH3 to Corophium thus derived is 3.1 mg/l (3.0–3.2). As shown in Table 2, Corophium is less sensitive to NH3 compared to other marine crustaceans. The LC50 value of 3.1 mg/l was used to calculate expected LC50 values (including confidence interval) for total ammonium at each pH, assuming toxicity is caused only by NH3. The results, as upper and lower level of the confidence interval, are plotted in Figure 3, together with the LC50 calculated from pH-controlled sediment experiments. The graph shows that all LC50 values from individual sediment tests with a pH of 8.3 or greater fit in the confidence range of the expected toxicity, whereas at lower pH the toxicity is higher than expected, indicating a combined NH +4 and NH3 toxicity at pH levels less than 8.3.

Table 2 LC50 (mg NH3/l) of several marine and estuarine crustaceans
Full size table
Figure 3
figure 3

NH3-only toxicity of Corophium volutator: solid lines give upper and lower level of the 95% confidence interval of the expected LC50 values, dots give the measured LC50 values, including their 95% confidence intervals.

Full size image

NH +4 toxicity at pH levels below 8.4 was also found when the ammonium responses of larvae of Macrobrachium rosenbergii were modeled (Armstrong et al. 1978). A contribution by the NH +4 ion has also been demonstrated for C. tentans and L. variegatus on the basis of the ratio between responses at the highest and lowest pH. The C. tentans test showed a more significant contribution by the NH +4 ion to the toxicity of total ammonium compared to L. variegatus (Schubauer-Berigan et al. 1995). For Artemia franciscana, total ammonium seems to be more important than NH3 in determining toxicity (Landau and Sanchez 1991), thus indicating a form of NH +4 toxicity.

Quality Criteria for Sediment Bioassays

Ammonium problems are relevant in sediment bioassays, especially those involving dredged materials. Based on pore water NH3 concentrations in dredged material, there is a 2% to 18% probability of seeing statistically significant toxicity in 10-day sediment toxicity tests with Leptocheirus plumulosus due to NH3 alone (Moore et al. 1997). Well-developed water quality criteria are needed for total ammonium concentrations in sediment bioassays. The results of this study show that water quality criteria for total ammonium concentrations in overlying water in sediment bioassays with C. volutator should be pH dependent. At a pH greater than 8.3, this criterion can justifiably be expressed as an NH3 concentration. At less than this pH, however, criteria should be expressed on a total ammonium basis, because of the combined NH +4 and NH3 toxicity. The data derived in this study can be used to propose pH-dependent water quality criteria for total ammonium in overlying water of a 10-day sediment bioassay with C. volutator.