- Top of page
- Materials and methods
A central purpose of environmental risk assessment is the protection of ecosystems from adverse impacts of chemicals. Because it is not possible to examine effects on ecosystems at large, single species, populations or communities are used to represent the ecosystem of interest. Micro- or mesocosms are used as model ecosystems under controlled conditions to simplify the studies of major ecosystem characteristics and observations are limited to a number of representative parameters. The observed effects are usually extrapolated to predict effects on natural ecosystems (Versteeg, Belanger & Carr 1999). Several approaches have been developed to reduce inherent uncertainty (Sijm et al. 2001). Predicted no-effect concentrations are commonly calculated using assessment factors (EC 2003).
Most of the data used in environmental risk assessment are based on the toxicological response at an individual or population level, including acute and chronic responses (Lin, Tokai & Nakanishi 2005), and are derived from single-species laboratory tests. Single-species tests are reproducible, but do not provide information on differences in species sensitivity or the effects of species interaction. The variation of sensitivities may be addressed by species sensitivity distributions (Kooijman 1987), which assume that variation in chemical sensitivity among species can be described by a statistical distribution (Roelofs et al. 2003). A species sensitivity distribution may be used to estimate the chemical concentration at which 95% of the species would be unaffected. This concentration is called 5% hazardous concentration (HC5) (Roelofs et al. 2003) and is used in probabilistic approaches to quantify environmental risks.
Further, effects on single species are usually recorded after exposure times of no more than one month, which is likely to be a shorter exposure time than that experienced by natural ecosystems. Therefore, chronic effects in ecosystems that might occur on a longer time scale are not addressed in these tests.
To assess variation between different exposure times, acute-to-chronic extrapolation, expressed as acute–chronic ratios (ACRs), have been proposed. The Technical Guidance Document (EC 2003) and ECETOC (European Centre for Ecotoxicology and Toxicology of Chemicals) aquatic toxicity data base define ACRs as ratios of an acute half-maximal effect concentration (EC50) vs. a chronic/subchronic no-effect concentration (NOEC) (Lange et al. 1998). Acute EC50 values usually refer to short-term exposure of invertebrates or fish for 48 h or 96 h, respectively. Chronic NOEC values usually originate from long-term exposures of invertebrates or fish for 21 days or 28 days, respectively.
Species interaction is an additional aspect of uncertainty when extrapolating from single-species tests to community- level effects. Contaminants may have the potential to change the outcome of species interactions and therefore influence community structure (Clements, Cherry & Cairns 1989), but species interaction is hardly ever considered in environmental risk assessment. One experimental approach to include effects caused by species interaction may be the use of community assays, representing one or more functional groups of an ecosystem. They possess a higher biological complexity with interacting species exposed simultaneously and all contributing to the observed response. Further, it is possible to monitor model ecosystems over prolonged time periods.
Species in a spatially structured environment, e.g. biofilms, are dependent on a complex web of symbiotic interactions (Hansen et al. 2007). They are characterized by competitive exclusion, predation or parasitism during succession but also by mutualism or commensalisms when coexisting in equilibrium. Additional stressors, e.g. toxicants may disturb this coexistence and induce a directed selection. Based on these considerations and according to the concept of pollution-induced community tolerance (PICT; Blanck & Wängberg 1988), it is possible to detect chronic effects from pollutants. The major mechanism behind PICT is the elimination of sensitive individuals or species and the proliferation of more tolerant ones during chronic exposure, meaning that the pollutant has exerted selection pressure on the sensitive members of the community (Nystrom, Bjornsater & Blanck 1999). Community tolerance can be examined by conducting several acute short-term tests. This offers the possibility to compare community responses to chronic exposure in terms of tolerance. Even more important, the observed effects include interactions within assemblages of various species, and therefore comprise several fundamental characteristics of natural communities that cannot be observed on a single-species level.
The scope of this present work was to compare commonly applied methods to calculate predicted no-effect concentrations in environmental risk assessment to effect-thresholds on community level. The objective was to account for species interaction as a factor of uncertainty in environmental risk assessment by applying the PICT concept. Periphyton was used to observe possible chronic effects over a period of several months. Tests were conducted using the phenylurea diuron, which provides a well-known mechanism of action and sufficient literature data on acute and chronic effects measured in single-species tests. It is a widely used herbicide, which acts as a photosystem II inhibitor (Devine, Duke & Fedtke 1993). This specific mechanism of action is universally effective in all autotrophic organisms and should therefore affect all species in periphyton communities; thus, differences in species sensitivity may derive from different uptake or transformation capabilities of the individual species. Environmental concentrations of diuron are usually below 0·1 µg L−1 but have exceeded 1 µg L−1 in surface near groundwater in Germany (LAWA 1997).
- Top of page
- Materials and methods
The pH value increased from minimally 7·1 in fresh river water to maximally 10·9 in aquaria at the end of 1 week. The mean pH value in fresh river water was 7·7 (SD 0·9) and in aquaria after 1 week 9·5 (SD 0·8). The increase in pH is due to the photosynthetic conversion of carbon dioxide reducing the carbonate content and therefore increasing pH. This phenomenon is well described in aquatic systems as well as for microbial mats (Revsbech et al. 1983). The microcosm experiment aimed to reconstruct standardized but semi-natural conditions; therefore, pH was not buffered and natural fluctuations of pH were accepted. Oxygen content ranged from 6·0–18·8 mg L−1 with a mean value of 9·1 mg L−1 (SD 1·9 mg L−1). Conductibility ranged from 269–757 µS with a mean conductibility of 466 µS (SD 124 µS).
To determine acute effects of diuron on photosynthesis, 20 short-term tests were conducted using periphyton cultivated without addition of diuron at different times during the long-term experiments of 3 to 12 weeks. As expected from its mechanism of action, diuron caused inhibition of photosynthesis at low concentrations in the µg L−1 range. EC50 ranged between 2·6 and 15·2 µg L−1 diuron with 80% of the values ranging between 4 and 9 µg L−1 diuron. According to Blanck (2002), this is termed the baseline tolerance of unexposed periphyton against diuron and will be used as reference for further results. The EC50 of 15·2 µg L−1 marks an outlier. The slope values range from 1·4 to 3·0 with a median value of 1·9. The determined LOEC was 1·2 µg L−1, the NOEC was 0·4 µg L−1.
Chronic effects of diuron were determined after long-term exposure over a period of several weeks. During the first experiment (I), several short-term inhibition tests were conducted at three different times to monitor tolerance development of periphyton. The EC50 values of controls were within the range of the baseline toxicity at all times and did not show any significant correlation with age or minimum fluorescence of the periphyton (data not shown). After 4 weeks, periphyton grown with diuron concentrations between 0–0·4 µg L−1 denoted no increases in tolerance compared to the baseline tolerance (Table 1). Diuron pre-exposure at 2 µg L−1 increased tolerance after 4 weeks up to an EC50 of 23·6 µg L−1. After 7 weeks of growth, an increased tolerance was determined for diuron concentrations between 0·08–2 µg L−1. The determined EC50 values show a strong correlation with the pre-exposure concentration of diuron (Fig. 1). After 11 weeks of growth, an increased tolerance was no longer determined for any diuron concentrations. These results show a minimal concentration of diuron to increase periphyton tolerance at 0·08 µg L−1 diuron during cultivation.
Table 1. Overview of EC50-values (µg L-1) in both experiments for all diuron concentrations applied during cultivation (n.a. not available; n.d. not determinable); enhanced EC50-values in comparison to baseline toxicity are highlighted; baseline toxicity (4-9 µg L-1) was calculated from short-term tests from control aquaria at different times throughout the experiment.
|Chronic diuron concentration (µg L-1)||EC50 values (µg L-1)|
|Experiment I||Experiment II|
|Controls (3-12 weeks)||4-9|
| ||after 4 weeks of growth||after 3 weeks of growth|
|0.0032||2.0 ± 0.3||n.a.|
|0.016||4.9 ± 0.4||n.a.|
|0.08||n.d.||4.4 ± 0.7|
|0.4||5.0 ± 1.3||3.2 ± 0.5|
|2||23.6 ± 6.3||4.7 ± 1.8|
|10||n.a.||3.0 ± 0.9|
| ||after 7 weeks of growth|| |
|0.0032||6.9 ± 0.6|| |
|0.016||8.3 ± 0.8|| |
|0.08||12.5 ± 1.1|| |
|0.4||12.6 ± 0.5|| |
|2||16.9 ± 3.5|| |
| ||after 11 weeks of growth||after 12 weeks of growth|
|0.0032||5.4 ± 0.4||n.a.|
|0.016||7.1 ± 0.6||n.a.|
|0.08||4.7 ± 0.7||21.4 ± 0.7|
|0.4||5.1 ± 1.4||16.6 ± 0.5|
|2||5.8 ± 0.8||22.8 ± 0.7|
|10||n.a.||26.0 ± 3.6|
Figure 1. EC50 concentrations of diuron after 1 h incubation as calculated from concentration-response curves of photosynthetic activity from 7-week-old periphyton, pre-exposed to diuron and solvent controls. ( Diuron pre-exposed periphyton – logarithmic fit y = ln(x − 0·4) R2 = 0·941).
Download figure to PowerPoint
The highest possible diuron concentration that would increase tolerance was assessed in the second experiment (II). After 3 weeks of growth, the EC50 values were determined for all concentrations, but no increase in tolerance was measured compared to baseline tolerance at this stage of community development. After 12 weeks of exposure, the tolerance increased for concentrations between 0·08–10 µg L−1 diuron up to EC50 values of 26 µg L−1 (Fig. 2).
Figure 2. Concentration-response curves of photosynthetic activity against diuron after 1 h incubation as measured with periphyton that was pre-exposed to diuron for 12 weeks and solvent controls from experiment II (curves 1 and 2: solvent controls; curve 3: 0·4 µg L−1; curve 4: 0·08 µg L−1; curve 5: 2 µg L−1; curve 6: 10 µg L−1 diuron).
Download figure to PowerPoint
For periphyton cultivated with 50 µg L−1 diuron, no tolerance could be determined at any time, as the data were not suitable for a logistic fit. Periphyton was severely disturbed; thus, a concentration of 50 µg L−1 diuron was not tolerable to periphyton. As shown in Table 1, diuron concentrations between 0·08–2 µg L−1 increased tolerance in both experiments.
In the second experiment, chronic effects of diuron were also monitored in terms of the development of chla-fluorescence and algal class composition using a multi-wavelength excitation PAM. Chla-fluorescence was similar in diuron-exposed and non-exposed aquaria until 5 weeks of growth. Thereafter, periphyton exposed to 0·08–2 µg L−1 diuron started to increase in minimum fluorescence much more than non-exposed periphyton for a further 2 weeks and then reached a steady state. The non-exposed periphyton started to increase in chla-fluorescence after 5 weeks, but reached a lower fluorescence level (examples shown in Fig. 3). Fluorescence of periphyton exposed to 10 µg L−1 diuron did not differ from non-exposed periphyton (Fig. 4). Exposure to 50 µg L−1 diuron caused a slight inhibition of periphyton fluorescence compared to non-exposed periphyton. Fluorescence patterns after 12 weeks for the whole dilution series is shown in Fig. 4.
Figure 3. Periphyton minimum fluorescence (F0) of Chla over time for two representative examples, measured with a multi-wavelength-excitation PAM-fluorometer ( periphyton with 0·4 µg L−1 diuron pre-exposure; non-exposed periphyton).
Download figure to PowerPoint
Figure 4. Total periphyton minimum fluorescence (F0) of chla for different pre-exposure concentrations of diuron, measured with a multi-wavelength-excitation PAM-fluorometer.
Download figure to PowerPoint
The algal class composition of periphyton was monitored over the whole period of 12 weeks. The non-exposed periphyton consisted of about 90% diatoms and 10% cyanobacteria (Fig. 5). The periphyton grown with low diuron concentrations between 0·08–2 µg L−1 started to differ from non-exposed periphyton after 8 weeks. The relative amount of diatoms decreased to about 50%, whereas the amount of cyanobacteria increased to about 40%. Only 0·08 µg L−1 diuron caused a change in the relative amount of green algae, which increased to about 20%. The high diuron concentrations of 10 and 50 µg L−1 did not cause clear changes in the algal class composition of periphyton biomass.
- Top of page
- Materials and methods
In the present study, acute and chronic effects of diuron on freshwater periphyton over a period of 3 months were observed in terms of photosynthesis inhibition, chla-fluorescence, community structure and tolerance development (PICT). Acute effects of diuron on photosynthesis inhibition, measured as variable chla-fluorescence in short-term tests, showed maximal effect concentration (EC50 values) of 4–9 µg L−1. This range included 80% of all the EC50 values determined in 20 independent short-term tests. EC50 concentrations for photosynthesis inhibition of marine microalgae communities, measured as incorporation of 14C, ranged between 4–20 µg L−1 diuron (Molander & Blanck 1992; Blanck & Dahl 1996; Arrhenius et al. 2004; Backhaus, Arrhenius & Blanck 2004) and 20–51 µg L−1 for stream periphyton (Dorigo et al. 2007). This consistency of the EC50 concentrations is in agreement with the specific mechanism of action of diuron and the high conservation of photosynthetic reaction centres within the plant kingdom (Bengtson Nash et al. 2005).
The observed chronic effects of diuron on periphyton included a clear increase of community tolerance for concentrations in long-term experiments between 0·08–10 µg L−1. Further, changes in algal class composition and increase of chla-fluorescence were observed for concentrations between 0·08–2 µg L−1. Minimum fluorescence was used as a proxy of chla in this study. An increase of the chla-content of algal cells after exposure to photosystem II inhibitors is a well-known phenomenon. Whether the observed increase in chla of exposed periphyton derived from a relative increase of chla per cell or of enhanced growth of algal cells remains unclear. From optical inspection, a higher density of periphyton cells seems plausible. The applied diuron concentration of 50 µg L−1 marks a highly toxic concentration that inhibited growth by almost 100%. The wide concentration range for the chronic exposure studies (0·0032–50 µg L−1) was chosen to cover the boundaries of induced tolerance. Environmental concentrations of diuron in surface near groundwater, reported by German water authorities (LAWA 1997) exceeded the PICT thresholds of this study, indicating that community shifts may occur in the environment.
However, differences between the two experiments were found regarding the time dependence of the observed PICT effect. The results originate from two independent but comparable long-term studies, which cultivated periphyton from a natural water source. Communities grown in natural water are subject to changes in water quality (e.g. nutrients) during their development. Therefore, the abundances and activities of different species, that might have different sensitivities, may change over time. As a result, in different studies the same point in time from the start of the experiment will probably resemble a different developmental stage of the community, depending on the composition of the inoculation. Additionally, detachment of large pieces of biofilm from the surface (sloughing) may be a confounding factor of biofilm succession, which relegates the biofilm to an initial succession phase. This will lead to varying EC50 values as found at similar times during the two experiments. Despite these variations, the effects caused by defined concentrations of diuron were highly reproducible, in terms of a certain concentration causing PICT or changes in biomass composition and abundance at any point in time during the study. Changes in composition of algal classes occur for the same diuron concentrations as the increase in community tolerance, clearly supporting the concept that more tolerant organisms will replace sensitive ones when a toxicant exerts selection pressure on the community. This was shown at a species level (Schmitt-Jansen & Altenburger 2005) and was detectable as a shift of algal classes.
Comparable studies were conducted by Molander & Blanck (1992) analysing PICT of marine periphyton for diuron and in combination with tri-n-butyl tin (TBT) (Molander et al. 1992). The threshold concentration causing PICT in both studies was around 10 µg L−1 diuron, which ranges about two orders of magnitude higher than the results of the present study. The process of replacement of sensitive species by less sensitive species due to selection pressure may take several reproduction cycles before PICT can be observed, despite the comparatively short generation time of unicellular algae and the fast action of diuron. Both studies by Molander & Blanck (1992) and Molander et al. (1992) showed that an observable PICT effect can be expected after 3–4 weeks when using a flow-through system, which allows new species to mix with the existing community at all times. However, the PICT tests of the present study conducted after 3 or 4 weeks, respectively, revealed no induced tolerance after 3 weeks, and after 4 weeks tolerance was only increased for 2 µg L−1 diuron pre-exposure. This may be due to the semi-static design of the experiments, leading to longer succession times than a flow-through system. In addition, it may be argued that 4 weeks might not be long enough to allow for the induction of tolerance by lower concentrations of diuron, as was found in this study after 7 weeks and 12 weeks of exposure, respectively. This means that diuron concentrations below the acute LOEC of 1·2 µg L−1 may need a longer period of exposure before PICT effects could be observed. Low PICT thresholds in the field were recently found by Dorigo et al. (2007), with a clear correlation of EC50 values for field periphyton and total pesticide concentration. Environmental concentrations of diuron ranged between 0·09 and 0·43 µg L−1 in that study, which was in the range the present study could detect PICT in microcosms.
Several other PICT studies have been conducted for heavy metals such as zinc, copper or arsenic, and triazines, such as atrazine and Irgarol 1051, or TBT (Table 2). For most of these chemicals, the EC50 concentrations in acute tests are about one to two orders of magnitude higher than the concentrations found to induce PICT in periphyton. In accordance with these, the present study revealed that for diuron, PICT appeared distinctly more sensitive than acute tests, in that the PICT threshold concentration was two orders of magnitude below the observed acute EC50 concentrations.
Because it is very cost- and labour-intensive to conduct community-level studies for every potentially hazardous chemical, extrapolation methods of chemical effects on communities from single-species results are needed for environmental risk assessments. They should consider specific community characteristics (e.g. variation in species sensitivity), which can differ among algal species by up to several orders of magnitude (Blanck 1984; Nystrom et al. 1999). Acute EC50 concentrations of diuron regarding photosynthesis inhibition range from 2–233 µg L−1 between different algal species (Holliste & Walsh 1973; El Jay et al. 1997; Schreiber et al. 2002; Bengtson Nash et al. 2005; Podola & Melkonian 2005). A popular method in environmental risk assessment is to estimate 5% hazardous concentration of chemicals from species sensitivity distribution. For diuron, the HC5 based on a log logistic distribution of EC50 concentrations regarding acute photosynthesis inhibition is 3·4 µg L−1 (Fig. 6). This value is in line with the EC50 values found for short-term exposure of communities in this study. The HC5 value based on a log logistic distribution of NOECs is 0·74 µg L−1, which ranges between the acute LOEC and NOEC found in the present study. At the same time, both these concentrations are distinctly higher than the threshold concentration of 0·08 µg L−1 diuron that induced tolerance. Consequently, the evaluation of species sensitivity variation using a statistic species sensitivity distribution may represent different species sensitivities, but does not sufficiently illustrate the effects observed in communities.
Figure 6. Species sensitivity distributions for diuron; data from single species algae assays were derived from the RIVM e-toxBase database (de Zwart, personal communication); included were only EC50, respectively NOEC, concentrations regarding acute photosynthesis inhibition.
Download figure to PowerPoint
Another factor that may be considered for extrapolation is variation in exposure time, such as acute-to-chronic ratios. A single-species test with diuron was conducted by Backhaus et al. (2004) using a unicellular green algae to assess the inhibition of reproduction after 24 h. The ACR calculated as the ratio of EC50 to NOEC was 9·4. Several data from other single-species assays also show ACR values around 10 (Isnard 1998; Roelofs et al. 2003; Ahlers et al. 2006). This consistency of ACRs indicates that in single-species assays, acute effects appear about one order of magnitude above the concentrations that typically cause chronic effects on single species. If effects from single species are to be extrapolated to communities, the question is whether these ACRs sufficiently describe chronic effects on the community level. This was tested for several photosystem II inhibitors (Table 2).
The calculated ACRs were applied to acute EC50 values (1–3 h photosynthesis inhibition) for algae, which resulted in the listed predicted chronic NOEC. In addition, the observed NOEC values are listed. All data are based on freshwater algae, except one PICT study for TBT. In this data set, the comparison of freshwater single-species tests and marine communities may confound the findings; however, with regard to the principal differences in test conditions between community and single-species studies, this factor may be of minor relevance. For atrazine and Irgarol 1051 the predicted and observed NOEC are in the same order of magnitude as concentrations that caused chronic effects on community level (PICT). In contrast, for TBT, copper and diuron the predicted and observed NOECs are at least one order of magnitude above concentrations that caused PICT. Therefore, an extrapolation from acute single-species effects to chronic effects on community level using acute-chronic ratios has its limitations.
In conclusion, neither variation in species sensitivities nor differences in exposure time could adequately predict the low threshold concentration of diuron that caused chronic effects on the community level. One factor that is not covered by species sensitivity variation or acute-to-chronic extrapolation is the interaction between species, which is an essential organizing force in communities. It is reported from a variety of autotrophic and heterotrophic biofilms that species interaction changes the coexistence of component species, community structure and function by resource competition or chemical interaction (e.g. allelopathic activity; Van der Grinten 2005; Hansen et al. 2007); thus, community tolerance can be regarded as a quantifiable surrogate of species interaction.
The results of the present study show that species interaction might account for differences between effect concentrations observed for single species vs. communities. It is shown that chronic community-level effects of diuron were not predictable from single-species tests. However, regulations such as the EC water framework directive (WFD) or the EC-REACH process rely on this type of information. Within the EC-WFD, Environmental Quality Standard (EQS) were defined for a set of priority pollutants and are now applied all over Europe. For diuron, an EQS of 0·2 µg L−1 is defined (EQS data sheet 13, Brussels 2005); however, this study has shown that this will not protect microalgal communities, indicating that water managers should consider higher-tier studies for the formulation of EQSs. Model ecosystems, e.g. micro- or mesocosms including several species of different sensitivity that interact with each other, but also PICT field studies (Dorigo et al. 2007) may provide a useful and complementary tool to current definition of EQSs. Alternatively, an additional safety factor could be used to account for species interaction effects at the community level and reduced if appropriate experimentations were done. Literature reviews revealed similar findings for other chemicals, indicating that species interaction might be a sensitive parameter for adverse effects on communities. Further PICT studies as a surrogate of species interaction should allow these findings to be applied and generalized to other chemicals. They may provide a prognostic tool for risk assessment of chemicals in the environment to reduce uncertainty when intending to extrapolate chemical effects from single-species tests to a community level in chemical hazard assessment.