Urban runoff can reach coastal aquatic environments; however, little is known about the effect of salinity on road runoff toxicity. The objective of this study is to investigate the toxicity of highway road dust over a salinity gradient from 5 to 35‰, in an estuarine benthic amphipod, Grandidierella japonica. Road dust toxicity was evaluated by assessing mortality after 10 days of exposure and short-term microbead ingestion activity of the amphipod. For all road dust samples considered, amphipod mortality increased with increasing salinity, whereas no significant difference in mortality was observed among test salinities in the reference river sediment. Ingestion activity during exposure to road dust decreased with increasing salinity. In fact, none of the individuals ingested any microbeads at salinity of 35‰. If assumed microbead ingestion is a proxy for feeding activity, high mortality at 35‰ could be attributed to aquatic exposure and not to dietary exposure. These findings suggest that road dust may have considerable impact on benthic organisms at high salinity levels.
Urban runoff contains different types of environmental pollutants, including metals and polycyclic aromatic hydrocarbons (PAHs) (Pitt et al. 1995; Sansalone & Buchberger 1997; Murakami et al. 2004). These pollutants discharge into receiving waters, causing adverse effects on aquatic ecosystems, particularly benthic organisms (Maltby et al. 1995). A large proportion of pollutants washes out in the early period of runoff; thus major toxicity could be attributed to the initial runoff, or first flush (Kayhanian et al. 2008).
A thorough understanding of the causes and factors affecting runoff toxicity is necessary to develop effective countermeasures and regulations to control runoff pollution. Some factors affecting runoff toxicity, such as photodegradation time, aeration time, screening of particle matters (Pitt et al. 1995), antecedent dry weather period, intensity and duration of rainfall (Greenstein et al. 2004), holding time (Pitt et al. 1995; Watanabe et al. 2011) and particle size distribution (Khanal et al. 2014), have been investigated. Although urban runoff flows into coastal aquatic environments with a wide range of salinities (Angelidis 1995), little is known about the effect of salinity of the receiving water on the runoff toxicity. Salinity affects mobility and bioavailability of pollutants. For example, increasing salinity from 0.5 to 31‰ caused higher elution of Co, Ni, Cu, Zn and Cd from several types of road dust (Schäfer et al. 2009). We speculated that salinity could significantly influence the runoff toxicity to aquatic and benthic organisms, with metal desorption promoted by increasing salinity and particularly in benthic environments.
It is also difficult to predict whether runoff input into saline waters actually enhances toxicity due to the following four reasons. First, dietary uptake of toxicants from solids is not negligible for toxicity to benthic organisms (Lee et al. 2000); thus higher desorption from solids might not always enhance toxicity. Second, changes in salinity results in changes in physiological processes of benthic organisms, such as dietary uptake rates (Normant & Lamprecht 2006). Third, increasing salinity generally reduces dissolved metals' toxicity to organisms (Hall & Anderson 1995), because free metal ion activity decreases due to complexation with chloride ions and competition with cations (e.g. Na+ and Ca2+) at the target sites of organisms (Di Toro et al. 2001). Finally, many potential toxicants other than metals are included in road dust (Rogge et al. 1993). Toxicity of some of them are affected by salinity in the opposite way to that of metals (Hall & Anderson 1995). Thus not only chemical analysis but also bioassays should be conducted to assess the effect of salinity on road runoff toxicity in benthic environments.
To date studies reporting the salinity effect on runoff toxicity using bioassays have not been conducted. Therefore we conducted a bioassay with an estuarine benthic species, Grandidierella japonica, to investigate the effect of salinity on the toxicity of road dust, as a component of the urban runoff solid phase.
Grandidierella japonica is known to present a great tolerance to salinity (Kikuchi & Matsumasa 1993; Santagata et al. 2009), and it has been reported from areas with widely fluctuating salinities (e.g. Greenstein & Tiefenthaler 1997; Nakashita et al. 2010), representing an ideal organism to study toxicity in a salinity gradient. The overall objective of this research was to investigate the effect of salinity on road runoff toxicity to benthic organisms. The specific objectives were: (i) to investigate 4-day mortality of G. japonica for aquatic copper as a reference chemical under various salinities (5–35‰); (ii) to evaluate the effect of salinity on road dust toxicity by 10-day mortality of G. japonica; and (iii) to examine whether dietary exposure of road dust was significant by microbead ingestion activity. The results are expected to contribute to the development of effective countermeasures and regulations to control road runoff pollution, particularly in estuarine environments.
MATERIALS AND METHODS
Source and pre-treatment of road dust samples and organisms
For this study, we used road dust samples collected and analysed in the previous research (Khanal et al. 2014). The concentrations of heavy metals and PAHs in road dust samples are shown in Table 1. Two road dust samples (A and B) were collected by highway sweeping vehicles. Sample C was collected from drainage pits in a similar highway. (Road dust A, B and C correspond to the samples from St. 6, St. 8 and St. 9 in Khanal et al. (2014), respectively.) The collected samples were sieved through 2 mm mesh, freeze dried and stored at 4 °C until use.
|Road dust Aa||Road dust A + Bb||Road dust Ca|
|Road dust Aa||Road dust A + Bb||Road dust Ca|
In this study, road dust samples tested were not diluted with uncontaminated sediment. A mixture of road dust A and B (3:1 w/w) was used instead of only road dust B, because inadequate quantities of road dust B were available to fully perform this experiment. The following pre-treatment of road dust samples was performed in both the 10-day mortality test and the microbead ingestion test: (i) road dust samples (200 g) were mixed with 775 mL of artificial seawater and adjusted to the test salinity in each test chamber; (ii) each test chamber was maintained at 25 °C and a photoperiod of 16 hours light to 8 hours dark (16L:8D) for 24 hours prior to beginning exposure; and (iii) half of the overlying water was replaced with fresh seawater 1 hour before adding 20 amphipods. The holding time was here fixed to 24 hours as described above, which followed the standard protocol for sediment toxicity test using estuarine amphipods (US EPA 1994). However, the holding time, or time elapsed between when road dust is mixed with water and when the organisms are exposed to the mixture, is known to affect road dust toxicity (Watanabe et al. 2011; Khanal et al. 2014).
Grandidierella japonica was collected from a tidal flat at the mouth of Nekozane River in Chiba, Japan, in 2009. The amphipods were cultured in an aquarium containing river sediment and artificial seawater (20‰ salinity) following the US EPA guidelines (US EPA 1994). The culturing aquarium was maintained at 25 °C and a photoperiod of 16L:8D. The river sediment was collected from a tidal flat from Arakawa River, and contained 19.5% ± 0.2% (mean ± standard deviation) of <63 μm fine particles and 3.7% ± 0.2% of organic matter (taken as equivalent to loss on ignition at 550 °C).
Four-day mortality test for reference toxicant
Amphipods can survive without additional food for 4 days, thus a 4-day mortality test has been conducted usually in order to evaluate aquatic toxicity (not including through dietary exposure). Conversely, a 10-day mortality test (described in the next section) has been developed for toxicity assay of contaminated sediment (US EPA 1994). In this research, prior to initiation of road dust toxicity tests we briefly investigated whether salinity really affects toxicity to G. japonica using the 4-day mortality test (US EPA 1994).
A solution of toxicants was obtained by adding stock solution of CuCl2 (1 mg Cu/L) to artificial seawater (20‰). The nominal concentrations tested were 0, 100 and 200 μg Cu/L. The amphipods were retrieved from the culturing aquarium and retained on a 0.5-mm mesh after filtering them through a 0.71-mm mesh. All the individuals were acclimated to the test salinity before the toxicity test (gradually adjusted by ±5‰/day; US EPA 1994). Twenty juveniles were transferred to each beaker, which contained 775 mL of seawater and artificial sediment (a mixture of 220 g of quartz sand and 10 g of kaolinite). Sixty individuals were used for each treatment (n = 3). During exposure to the toxicants, each test chamber was held at 25 °C and a photoperiod of 16L:8D. The overlying water was not aerated and no food was supplied. The number of surviving individuals was counted at the end of the test. Missing amphipods were considered dead.
Ten-day mortality test for road dust
Instead of a 4-day test, a 10-day mortality test was conducted to evaluate road dust toxicity under various salinity levels (ranging from 5 to 35‰). The following three changes were made to the procedure of the 4-day mortality test: (i) exposure duration for the mortality test was extended to 10 days (US EPA 1994), and then each test chamber was supplied with a gentle ﬂow of air throughout the test to maintain the levels of dissolved oxygen at over 60% of saturation, and with 20 mg of TetraMin® (Tetra Japan, Tokyo, Japan) on alternate days; (ii) road dust samples prepared as described above were used; and (iii) the river sediment used for culturing was used as reference.
Salinity was measured daily using a conductivity meter (CD-4307SD, Mothertool, Japan), and adjusted to the test concentration ±0.5‰. The concentration of dissolved oxygen was also monitored daily with an oxygen meter (DO-31P, DKK-TOA, Japan). Ammonia concentration and pH in the overlying water were measured at the start and end of the toxicity test. Total ammonia concentration was determined using a Hach kit (TNT™ 830 Ultra Low Range; 0.015–2.00 mg/L NH3-N) at 690 nm. Only if these measured values were within the tolerance limits of G. japonica (dissolved oxygen: above 60% of saturation; pH: 7.5–8.5; total ammonia: below 30 mg/L) were the results used for further analysis (US EPA 1994; Lee et al. 2005).
Microbead ingestion test for road dust
Besides a traditional sediment toxicity test whose endpoint was mortality, a microbead ingestion test was conducted. Ingestion of suspended microbeads could be a proxy for dietary uptake, because G. japonica is a filter feeder (King et al. 2006) which captures suspended particles.
The microbead ingestion test used was a modification from a previously described method applied to Daphnia magna (De Coen & Janssen 1998). Ingestion activity during exposure was evaluated using fluorescent polystyrene microbeads (2.5–4.5-μm diameter, Spherotech, Lake Forest, IL, USA). The test chambers and amphipods were prepared following the procedure described as pre-treatment for road dust. Sixty minutes after inoculating the solution with the amphipods, a suspension of fluorescent microbeads was added to the overlying water (25 mg/L). The amphipods were allowed to graze on the fluorescent particles for 30 minutes. Subsequently, they were fixed with 70% ethanol. In a preliminary test, it was demonstrated that all amphipods ingested the microbeads when left to graze for 30 minutes under control conditions. The microbeads attached to the body surface were removed by gently washing the amphipods with 5% Tween® 80 and Milli-Q water. We checked the samples using a fluorescence microscope, to ensure that almost all the microbeads attached to the body surface were washed out. The washed amphipods were individually homogenized in Milli-Q water by ultrasonication. The resulting liquid was then filtered through a glass fibre filter (0.6-μm pore size). The microbeads retained on the filter were counted using a fluorescence microscope. Data were analysed using R software (ver. 3.0.2; R Development Core Team 2013, Vienna, Austria).
The water-soluble fraction of road dust constituted major toxicity to a freshwater ostracod (Watanabe et al. 2011). We also estimated the contribution of dietary exposure to toxicity under various salinities using the microbead ingestion test to determine if salinity might change which fraction of road dust is more toxic.
RESULTS AND DISCUSSION
Salinity effect on copper toxicity
There was no statistical difference in mortality among salinities for each copper concentration (Figure 1; analysis of variance (ANOVA), p > 0.1). The effect of salinity on 4-day mortality of G. japonica for dissolved copper was not significant.
The reason why the effect of salinity could not be observed was not technical failure, but just because the effect of salinity was not sufficient to be observed within the range of copper concentration tested. The results of the 4-day mortality test were reliable, because the level of observed mortality was comparable with King et al. (2006), in which 4-day NOEC (no observed effect concentration) and LOEC (lowest observed effect concentration) of G. japonica for copper at 30‰ salinity was 90 and 190 μg Cu/L. In this study, 4-day NOEC and LOEC for dissolved copper at 35‰ was 100 and 200 μg Cu/L, respectively (Dunnett test, p < 0.05) (Figure 1).
Salinity effect on road dust toxicity
Increasing salinity from 5 to 35‰ caused an increase in amphipod mortality for all three road dust samples (Figure 2). This effect of salinity on road dust 10-day mortality was more apparent than that on copper 4-day mortality (Figure 1), although exposure time was different.
For the river sediment, there was no significant difference (ANOVA, p > 0.1) in mortality among salinities. This result was consistent with previous reports showing that G. japonica maintained a constant haemolymphosmotic pressure under widely fluctuating salinities (Kikuchi & Matsumasa 1993). In addition, it has been reported that G. japonica inhabits areas where the salinity is 25–35‰ but sometimes declining below 10‰ (Greenstein & Tiefenthaler 1997; Yamaguchi et al. 2000; Nakashita et al. 2010). The high ability of G. japonica to adapt to fluctuations in salinity highlights the fact that the range of salinities used in this study (in particular 25–35‰) is not unrealistic for this species.
The causative toxicants in road dust could not be identified within the frame of this study due to the following reasons. First, only a few LC50 values for G. japonica were available; we could not determine whether the target toxicant significantly contributed to the observed toxicity. Second, correlation analysis between the content of all considered toxicants and amphipod mortality at each salinity level in this study (Supplementary Material, Figure S1, available online at http://www.iwaponline.com/wst/072/304.pdf) was not sufficient to find the major candidate toxicants in the road dust samples, although the number of road dust samples was limited. In fact, road dust toxicity is different depending on samples, suggested by Khanal et al. (2014) in which road dust toxicity to an ostracod had no significant correlation with the content of heavy metals and PAHs.
Here we briefly discuss three possible reasons behind the increase in road dust toxicity with the increase in salinity. First, an increased bioavailability of the metals could explain this relation. An increase in salinity could result in a higher level of elution of heavy metals from road dust (Schäfer et al. 2009). Thus, higher toxicity might be induced at high salinity levels, if the dissolved metals are more bioavailable compared with those bound to the dust particles.
Second, toxicants other than metals can also explain the increase in road dust toxicity by increasing salinity. For instance, anionic surfactants, which have been reported to be one of the causes of runoff toxicity (Kayhanian et al. 2008), become more toxic to aquatic organisms in hard water than in soft water (Tovell et al. 1974). The toxicity of anionic surfactants is assumed to increase at high salinity levels, although the effect of salinity on anionic surfactant toxicity has not been reported.
Third, disturbances to the osmoregulatory function, which enables the organism to adapt to different salinities, may also explain the relationship between salinity and toxicity observed here. For example, osmo- and iono-regulation in estuarine and marine crustaceans are disrupted by exposure to heavy metals, resulting from gill damage (Bianchini & Carvalho de Castilho 1999). To verify these possibilities (i.e. to identify the cause of the increase in toxicity), further chemical analyses and molecular bioassays are recommended.
Microbead ingestion in road dust
After microbead ingestion tests, the digestive tract became coloured with yellow fluorescence in some individuals (Figure 3(a)). The data on ingested microbeads were non-normally distributed and heteroscedastic among salinities (Shapiro–Wilk test, Bartlett's test, p < 0.05); therefore, percentile values were used in Figure 4.
Ingestion activity in the road dust mixture significantly decreased with increasing salinity (Kruskal–Wallis test, p < 0.05; Figure 4), although that in the river sediment was not the same among salinity levels (Kruskal–Wallis test, p < 0.05). None of the individuals exposed to road dust for 30 minutes ingested any microbeads at 35‰ salinity according to microscopic observations (Figure 3(b)).
Considering microbead ingestion as a proxy for feeding activity, extremely high mortality at 35‰ after 10 days exposure could be attributed to aquatic exposure and not to dietary exposure. Although the test organisms may recover their ingestion activity during 10 days (De Coen & Janssen 1998), the aquatic toxicity of road dust at 35‰ was considered to be larger compared with that at lower salinities. One possible explanation of high aquatic toxicity at higher salinities is a larger elution of heavy metals as described in the section ‘Salinity effect on road dust toxicity’. In contrast to at 35‰, we cannot determine whether the exposure route was dominant at lower salinities.
Ingestion activity largely varied among individuals at each salinity level (Figure 4), preventing accurate toxicity measures for road dust (e.g. inhibition rate). This large variation may relate to individual differences than to heterogeneity of the sediment samples, because this high variation was observed even in an artificial sediment (Figure S2, available online at http://www.iwaponline.com/wst/072/304.pdf; coefficient of variation values were above 40%). In preliminary research we studied the experimental conditions for the microbead ingestion test, such as exposure time and diameter of microbead, in order to reduce the variation; however, the variation in ingestion activity was still large. It will be essential to investigate further the experimental conditions. For example, Agostinho et al. (2012) demonstrated that small test vials were desirable in minimizing the variation observed in nauplii ingestion of amphipods.
The exposure to the road dust sample enhanced microbead ingestion compared with that of the control condition except at 35‰ (Figure 4), although it caused a relatively high level of mortality (Figure 2). We propose two possible explanations for this discrepancy. First, the road dust mixture would contain more nutrients for the amphipods than the river sediment. We observed that the number of ingested microbeads increased with increasing amount of added TetraMin® during exposure (Figure S2). Although the organic matter content did not differ between samples (road dust mixture: 4.1–4.8% by Khanal et al. (2014); river sediment: 3.7%), how organic matter content relates to the quantity of food available and the level of food selectivity in G. japonica is not clear. Hormesis could also explain the discrepancy between the ingestion and mortality data. Hormesis is a phenomenon whereby low level of exposure to stressors can stimulate the response of organisms (Calabrese 2010). The exposure to road dust under lower salinity conditions was less toxic to the amphipods than that under higher salinity, as indicated by the mortality after 10 days, and thus might stimulate ingestion activity.
Suggestions for runoff pollution management
Our results demonstrated that increasing salinity from 5 to 35‰ resulted in high mortality of the estuarine amphipod for highway road dust. This finding suggests that the salinity level in receiving waters is an important factor for road runoff toxicity, and that selection of discharge outlet locations should be paid attention. In terms of road runoff pollution, environmental conditions in receiving waters, such as the salinity level, have been almost overlooked. Many management practices have focused on source control (e.g. street sweeping) and stormwater treatment until discharge into receiving waters (e.g. construction of retention ponds) (Loganathan et al. 2013). Therefore our results provide a new viewpoint for management of runoff pollution.
While undiluted road dust samples were used in this research, road dust would be diluted by natural sediment in real environments. However, almost all dust deposited on the road surface discharges into receiving waters within the first few minutes of runoff (Sanalone & Buchberger 1997), which suggests that road dust is not distributed uniformly and accumulates in local hot spots. Our results suggest that environmental assessment should take into account salinity levels in such local hot spots in order to reduce the toxicity of road runoff.
Three types of toxicity tests were performed with an aim of investigating the effect of salinity on road dust toxicity to the estuarine benthic amphipod G. japonica. The results obtained in this study could contribute to the development of road runoff management, particularly in estuarine environments.
The effect of salinity from 5 to 35‰ on 4-day mortality of G. japonica for dissolved copper (100 μgCu/L, 200 μgCu/L) as a reference toxicant was not significant.
Increasing salinity from 5 to 35‰ caused higher amphipod mortality after 10 days for all road dust samples tested. This effect of salinity on road dust toxicity was more apparent than that on copper 4-day mortality. Increasing salinity had no significant effect on mortality in the river sediment, which suggests that high levels of salinity itself were not toxic to the amphipod. We could not determine the causative toxicants in road dust, but proposed three possible reasons for increasing road dust mortality with increasing salinity.
Ingestion activity in the amphipods exposed to road dust decreased with increasing salinity. In particular, none of the tested organisms ingested any microbeads at 35‰ salinity. If microbead ingestion is assumed as a proxy for feeding activity, high mortality at 35‰ could be attributed to aquatic exposure and not to dietary exposure.
We would like to thank Dr H. Watanabe (National Institute for Environmental Studies, Japan) for collecting and culturing the test organisms. We also wish to thank Prof. H. Furumai, Mr S. Karasawa, Dr R. Khanal and Mr K. Tsukahara (The University of Tokyo) for their help in sample collection. We would like to thank the anonymous reviewers for their helpful comments on an earlier version of this paper. The present study was financially supported by Grants-in-Aid for Scientific Research (B)(24360213), Japan Society for the Promotion of Science (JSPS) and the Steel Foundation for Environmental Protection Technology (SEPT), Japan.