2. "Known unknowns” – unregulated micropollutants and chemical mixtures

2.1.         Introduction
Under the WFD, surface water assessment is separated into chemical and ecological status. Such separation may reflect a practical solution for water regulation but it is artificial for the environment. This chapter considers ways to gain evidence for better linking chemical and ecological status of surface waters in future.

Following the reduction of gross pollution, considerable effort in recent years has been put into developing ways to assess the impact of chemicals from an organism’s perspective i.e. “what concentrations of which substances affect the healthy functioning of an ecosystem?” A better understanding could allow improved targeting of measures to reduce harmful concentrations of pollutants. Alongside this, concerns have grown about the “cocktail effect” – mixtures of low concentration chemicals which in combination may cause harm. Some of the challenges and proposed solutions towards improving assessment of chemical risks in water are considered below.

comments (1)

2.2.         Chemical and ecological status

The Water Framework Directive (WFD) assesses chemical and ecological status of surface water bodies separately. However, organisms living in the water experience an integration of all the influences present. The different statuses can lead to the criticism that the reported “chemical status” may be remote from what is actually occurring in the water ecosystem.

The chemical status of surface waters under the WFD is based on a comparison of measured concentrations of EU-wide consented priority substances with target levels established under the Environmental Quality Standards Directive (EC, 2008a). Ecological status is assessed from monitoring data on biological quality elements (BQE) such as benthic invertebrate fauna, phytoplankton, macrophytes, and fish. Additionally, data on hydromorphology (physical characteristics), physico-chemical water parameters and RBSPs can be used (figure 2.1). Owing to the particular geographic circumstances of any particular water body, assessment of ecological status is made with reference to specific local factors.

 Figure 2.1: Overview on the current status assessment approach under the WFD

The value of chemical measurements in rivers and lakes is that they allow direct comparison of concentrations between sites. Furthermore, they can be related to emission loads and, therefore, controls can be directed towards specific sources of chemical pollution. However, among the criticisms of this approach are that ecological structures and functions, key targets of chemical pollution, can be poorly related to specific chemical measurements. In particular, pollution by emerging compounds may be overlooked.

Efforts to link chemical occurrence and ecological effects are not required under the WFD and failures to achieve good ecological status, solely driven by chemical pollution (e.g. RBSPs), are rarely observed.

comments (2)

Fig 2.2a-d shows chemical status with and without uPBTs, as well as the ecological status, by country.

Figure 2.2a shows chemical status by country (EEA, 2018a). A number of countries have reported 100% failure of chemical status owing mainly to pollution by mercury. The 2013 Priority Substances Directive (EU, 2013b) identified 4 groups of substances as “ubiquitous, persistent, bioaccumulative and toxic” (uPBT) (section 1.2). Omitting these from the calculation of chemical status increased overall good chemical status to 81% ((graph C). Meanwhile, ecological status is shown in graph B. 

Figure 2.2a: Chemical status in surface waters, with uPBTs


Figure 2.2b: Chemical status in surface waters, without uPBTs


 Figure 2.2c: Ecological status in surface waters


 Figure 2.2d: Ecological status of River Basin Specific Pollutants




It is difficult to see what relationship, if any, exists between figures 2.1A-C. It can also be seen that in many water bodies, the RBSPs have not been reported in the assessment of ecological status (Fig 2.1D). 

Information on whether and to what extent chemical and ecological status indicators are correlated has the potential to be used to indicate the effects of pressures and, potentially, explain causes of observed ecological effects, providing evidence for decision-makers. The scientific community has proposed diagnostic approaches to unravel links between ecological effects and chemical contamination, and strong interest in this research has been indicated by stakeholders of water management (Brack et al. 2015) (Box 2.1).

comments (6)

Box 2.1

Box 2.1 SOLUTIONS – pollutants in land and water management

This EU FP 7 project assessed how existing WFD practice could be brought more up-to date with currently available scientific knowledge (Brack et al. 2017). Recommendations included:

·       use of effect-based methods for pollution investigation and assessment

·       use of passive sampling for bioaccumulative pollutants

·       integrated strategy for prioritisation of contaminants in monitoring

·       consideration of priority mixtures of chemicals

·       historical burdens accumulated in sediments

·       models to fill data gaps

·       tiered approach in investigative monitoring to identify key toxicants


comments (0)

2.3.         Evidence for chemical pollution causing ecological effects

The established way of identifying clear links between a chemical and its effect on organisms is through concentration-response relationships, for example by comparing an organism’s health response with increasing concentrations of a chemical. As it is impossible to assess the sensitivity of all organisms to all pollutants, assessment factors are applied to accommodate for uncertainties and data gaps, including chronic effects. Where an EQS has not been established for a substance, experimentally-derived effect concentrations may be compared with estimated or measured environmental concentrations (figure 2.2).

Figure 2.2: Individual chemical risk assessment is based on comparison of single chemicals concentrations in the environment with standards


comments (1)

Box 2.2 Definitions

Acute toxicity – adverse effect on an organism after short-term exposure.

Chronic toxicity – adverse long-term effect after long-term exposure (typically at lower concentrations than those causing acute toxicity).

Mixture toxicity – adverse combined effect after exposure to multiple pollutants

Mode of action – understanding of how a chemical acts in an organism or ecosystem

Bioassay – biological test system (organism or cells)

Effect based method (EBM) – bioassay suitable for environmental monitoring 

Molecular target  - biomolecule (e.g. protein) that directly interacts / binds with a chemical

comments (2)

A pioneering study by Malaj et al. (2014) used monitoring data on chemical concentrations, based on data reported in WISE–SoE. The authors considered more than 200 substances monitored in European freshwater systems. They reported an acute risk at 14% and a chronic risk at 42% of the sites investigated (figure 2.4 A). One issue identified using this approach, however, is that the expected risk increases with the availability of chemical monitoring data. Where concentrations are unknown, they cannot be used in the assessment and so this may result in a skewed result, with sites for which information is available appearing worse than those for which this information is not provided (figure 2.4 B). A further issue is that the availability of data for acute toxicity is much greater than that for chronic toxicity, meaning that the chronic risk assessment is more dependent on assessment factors and thus prone to larger errors.

Figure 2.4: A) Acute and chronic risk estimates for European water bodies based on reported chemical monitoring data and calculated using risk estimates for individual compounds;  


Fig 2.4 B) Correlation between chemical risk and number of chemicals analysed for acute risk (ART = acute risk threshold, CRT = chronic risk threshold); figures from Malaj et al. 2014)


Recent research indicates that chemicals contribute to a significant but varying extent to the total effective stress in river ecosystems (Schäfer et al. 2016, Rico et al. 2017). Rico and colleagues (2016) showed that variation in invertebrate communities could be mainly explained by habitat and water quality, with physico-chemical parameters (e.g. dissolved oxygen) explained more of the variation than metals or organic contaminants. The authors reported that it was difficult to find direct links between individual contaminants and ecological effects.

In the EEA's RBMP Assessment (EEA, 2018a), it is highlighted that countries with good ecological status for benthic invertebrates also have lower levels of pressures. This seems true especially for diffuse pollution and hydromorphological pressures. To identify e.g. pressure-related failures of good ecological and chemical status might require a second line of assessment, beyond the prevailing basic one-out all-out principle. Such studies could be successful with pollutant concentrations instead of EQS exceedances and organism compositions instead of biological quality element classes.

In conclusion, it is rarely possible to explain observed effects in ecosystems based on knowledge about the presence of individual chemicals, while ecological impact information alone is similarly not sufficient to identify the chemicals causing that impact. Instead multiple lines of evidence are needed.

comments (2)

2.4.         Dealing with mixtures of chemicals

For establishing causal relationships between chemical pollution and ecological effects, it has to be appreciated that in the real world there are no cases where only a single substance occurs in the environment. Emissions data and research show that the aquatic environment has to deal with mixtures of chemicals, which contain many more substances than just the priority substances. Nutrients from urban point sources, agricultural diffuse pollution, metals from stormwaters from cities and atmospheric deposition, as well as many potentially harmful organic chemicals from urban waste water and agriculture, have been shown to be present in freshwater systems simultaneously. Indeed, scientific monitoring approaches highlighted the co-occurrence of hundreds of chemicals in different freshwater bodies (e.g. Loos et al. 2009 & 2013, Moschet et al. 2014). This complexity mismatches with the single substance approach of current chemicals assessment under the WFD.

The occurrence of chemical mixtures in freshwater systems is the result of different sources and different patterns in time, space and concentration (e.g. Baker & Kasprzyk-Hordern 2013, Beckers et al. 2018) and so does the respective risk for the ecosystems. The challenge is to figure out which of the many substances present are most important for the toxicity of a mixture.

Efforts exist to simplify this complicated picture. In essence, these aim to separate and categorize the issues of pollution, impact and identification of key chemicals to achieve a problem-targeted assessment (figure 2.5). Statistical methods are used to characterise complex pollution situations and relate these to sources (Posthuma et al. 2017). This approach offers the potential for identifying categories of mixture: “typical” i.e commonly-occurring, or “priority” i.e. containing substances which are of particular concern in a mixture, for instance because they promote toxicity. This is particularly relevant for the diverse and numerous organic micropollutants for which single representative candidates on lists of regulated substances are often outdated or may be substituted by substances with potentially similar toxicity when regulation comes into play. The combined action of similar compounds occurring together is not captured at all (Altenburger et al. 2015).

Examples for co-occurrence of similar compounds comprise the neonicotinoid insecticides imidacloprid, thiacloprid, acetamiprid which have been shown to occur simultaneously in water bodies but also antibiotic drugs such as azithromycin, erythromycin, and clarithromycin or the herbicides (e.g. diuron and isoproturon).

Figure 2.5: Dealing with mixtures in water management through differentiation into pollution (priority mixtures), effect (impact of mixtures) and risk (drivers of mixture toxicity) issues (modified from Altenburger et al. 2015)

A study by Busch et al. (2016) described the diversity of potential molecular targets for contaminant-biosystem interactions. In this study 426 organic chemicals were summarized to be detected in European freshwaters, containing 173 pesticides, 128 pharmaceuticals, 69 industrial chemicals and 56 other compounds. These targeted more than 100 different biological molecules known to exist in aquatic organisms. This complicated picture was simplified by considering the ways in which the chemicals acted upon organisms – “modes-of-action”. 30 mode-of-action categories were identified for freshwater contaminants (figure 2.6), so that even with a potentially unlimited number of chemicals, there is a limited range of adverse biological effects. This approach could be used to simplify toxicity assessment.

Figure 2.6 Modes of biological action of organic micropollutants in European freshwaters

 Source: Busch et al. 2016)

Notes: Abbreviations: GABA – gamma-Aminobutyric acid (chief inhibitory neurotransmitter in the mammalian central nervous system); nAChR - nicotinic acetylcholine receptor (see table 3.1); ATP – adenosine triphosphate (energy carrier in the cells of all known organisms), DPP4 - Dipeptidyl peptidase-4 (an enzyme)


comments (4)

The largest group of organic micropollutants with a known mode of action identified in this study were neuroactive compounds, which affect or interact directly with the nervous system.

Chemicals that affect the nervous system interact with different molecular targets, e.g. different insecticides binding either to the nicotinic acetylcholine receptor or inhibiting the enzyme named acetylcholine esterase (table 2.1). Both affect the signalling in the nervous system and mixtures of such chemicals will enhance the effects. Aquatic invertebrates might be particularly at risk owing to exposure to mixtures of different kinds of insecticides, while other species, such as fish, might be affected by the presence of antidepressant or antiepileptic pharmaceuticals that affect the nervous system of fish, possibly in combination with effects caused by insecticides.  This means that chemicals, such as pesticides and pharmaceuticals, which are intended to act via certain modes of action in a certain species, can affect other species as well. For industrial chemicals, such as bisphenol A, PAHs and pBDEs, it is rather difficult to define a specific mode of toxicological action as those can show complex and multiple modes of action. They have been found to cause different chronically relevant responses, indicating long term toxicity such as endocrine disruption and mutagenicity, across various organisms including humans.

Table 2.1: Examples for mode-of-action categories and related mechanisms of chemical action (For further details see Busch et al. 2016)

comments (0)

It can be difficult to predict the outcome of chemical mixtures on biological effects. In broad terms, the chemicals might a) act independently of each other, exhibiting individual toxicity; b) in combination, be more toxic, as a summed total of the individual chemicals or more toxic than that; c) be less toxic as the chemicals interfere with each other in toxicity mechanisms. For chemicals in a mixture that have the same mode of action, an additive combination effect may be expected (Altenburger et al. 2015, Figure 2.7). Developing knowledge in this way, considering effect contributions from all compounds detected, would be expected to provide stronger association between chemical and ecological assessments.

Figure 2.7 Predicting the outcomes of mixtures - concentration addition for compounds with the same mode of action.

comments (1)

2.5.         Examples combining chemical and biological monitoring

While modern effect-based methods have been proposed for mixture assessment, as a complement to chemical and ecological monitoring, precedent already exists in this respect. Such methods offer something similar to the “biological oxygen demand” (BOD) which measures an overall condition in the water while not specifying the cause. Despite this lack of specificity, BOD is widely used in water management to protect surface waters (EEC, 1991; EC, 2000).

Currently, there are few requirements to use effect-based information in regulatory assessment. An example where effect-based monitoring is used for assessment is the Marine Strategy Framework Directive (EC, 2008b). Different descriptors of good environmental status, such as “concentrations of contaminants at levels not giving rise to pollution effects”, are defined and the assessment allows the integration of data on biological effects (Lyons et al. 2017). The application of bioassays for measuring the occurrence of dioxins and PCBs in foodstuffs (EU 2012) demonstrates how effect-based assessment might operate in a regulatory framework. The value of such information is that it integrates the effect of mixtures of chemicals irrespective of whether the combined effects are additive or different from an expectation.

For example, the total potency of compounds with estrogenic activity in a water sample can be determined by measuring the activity of the estrogen receptor in laboratory in vitro assays. Ideally, the bioassay captures the total effect of all chemicals with estrogenic effect in a sample. Practically, difficulties exist, though the robustness of techniques has improved for some modes of action in recent years (e.g. Altenburger et al. 2018, Leusch et al. 2018, Kunz et al. 2017).

For regulatory monitoring, techniques need to be robust and reliable, to meet legal challenge and ensure that investments are based on sound evidence. A series of International Standards Organisation (ISO) standardized methods is available for the use of biological methods for the assessment of effluents on water quality[1]. The EU water directives transposed into national regulation allow Member States to set requirements appropriate for the country level e.g. the German “Abwasserverordnung” (AbwV, 1997) specifies standard methods for specified types of waste waters.

To demonstrate the application of biological effect tools in monitoring, case studies have been undertaken. In a pilot study by Escher et al. (2014) the efficacy of different waste water treatments was determined using the observable effects of enriched water samples in about 100 different miniaturized and mainly cell-based bioassays (figure 2.8). Results showed the presence of different chemicals at different levels of pollution with diverse modes of action.

Figure 2.8: Examples of organism and cell-based bioassays for water monitoring

In a case study performed within the European FP7 project SOLUTIONS, Neale and colleagues (2017) investigated the WWTP effluent, upstream, and downstream river water samples in Switzerland. They compared bioanalytical results from 13 bioassays with results from chemical analysis of 405 compounds (see figure 2.9A).

Figure 2.9: Example of a comparative analysis of chemicals and combined effects using component-based mixture predictions (taken from Neale et al. 2017)



They found that of the 10 detected herbicides known to inhibit the photosystem II (PSII), terbuthylazine and diuron could explain the majority of biological effects (figure 2.9 B). The authors also showed that the detected chemicals could explain between 45 and 108 % of the observed biological effects. In samples collected upstream of the WWTP, only a fraction of the total measured effect could be explained by the detected chemicals.

[1] https://www.iso.org/committee/52972/x/catalogue/

comments (5)

2.6.         Towards effect-based methods in monitoring and assessment

Assessment under the WFD currently does not consider mixture effects. It is, therefore, possible that concentrations of priority substances could be slightly below their respective EQS, meeting good chemical status, while the actual combination of substances present could be harmful. For example, if all five PSII inhibitors from the priority substances list were detected, individual concentrations might be at good chemical status but the mixture could nevertheless cause adverse effects (Figure 2.6). Additionally, while the list of priority substances represents certain hazardous chemicals, there are other substances present in surface waters which could contribute to mixture effects.

Mixture effect considerations could thus be integrated into the existing assessment schemes following two approaches that could be anticipated:

Compound-based mixture prediction: EQS for mixtures of similar acting compounds could be established and potentially considered in chemical status assessment.

For example, an EQS for the sum of all six PSII-inhibitors could be defined as the sum of the single substance concentrations divided by the single substance EQS. If this sum exceeds “one”, than the EQS of priority PSII-inhibitors is exceeded.

Mixture effect detection using effect-based methods: Joint effects measured with a bioassay instead or in addition to single chemical compound concentrations might be considered as indicators for the ecological status assessment.

For example, instead of determining the concentrations of each PSII inhibitor in a water sample, the sample would be concentrated and tested in a dilution series using a bioassay (e.g. algae growth inhibition test). At the point where the toxicity ceases, the dilution factor would be compared with the test result of a defined reference compound (e.g. diuron).

Currently, several whole organism-based assays and some cell-based assays are ready for routine use in an effect-based monitoring. Readiness for use implies fulfilling requirements regarding standardization, robustness, reproducibility - for several modes of action, we lack specific bioassays, even though there are many techniques available to researchers. Within the WFD water quality assessment, selection of the relevant bioassay could be derived from the biological quality elements assessed in the water body. Organism-based bioassays therefore could support the link between chemical and ecological monitoring and assessment (figure 2.10).

Figure 2.10: Biological effect assessment could serve to close the gap between ecological and chemical assessments and gain causal relationships

The European Commission (Wernersson et al. 2015) gives a summary of available bioanalytical tools in the technical report on aquatic effect-based monitoring tools under the WFD. Their readiness for monitoring applications has been evaluated in several projects (e.g. Kienle et al. 2015). These tools can be applied and used in a modular manner depending and targeted on the desired level of evidence (figure 2.11).

Two applications of effect-based methods can be foreseen: 

The monitoring of chemical impact on biological quality elements (BQEs):

For effect monitoring, a module comprising different organism-based bioassays representing the different BQEs would provide evidence for total chemical impact. It would also enable direct linkage of effect observations with ecological monitoring data (figure 2.9 A, figure 2.10). However, to detect chemicals with an impact that emerge over a longer time scale, such as endocrine disruptors or mutagenic and genotoxic compounds, additional bioassays, such as cell-based mutagenicity assays and estrogen receptor activation assays should also be implemented (figure 2.9 B).

Investigations of pollutants which cause effects:

When investigating chemicals which could be causing effects through specific modes of action (table 2.1) or on specific, stress-related endpoints, additional bioassays are available. The application of such in vitro detectors may also be used to protect specific uses of a water body, e.g. drinking water abstraction.

Figure 2.11: Modular approach for application of bioassays in monitoring

comments (13)

2.7.         Challenges

The implementation of effect-based methods into monitoring routines or diagnostic screening approaches would require agreement on the bioassays to be used. Robust bioassays have been developed for some organisms (such as invertebrates like Daphnia) and assays for the detection of estrogenic compounds, with detailed recommendations for application in monitoring (e.g. Kunz et al. 2017).

Broadening the use of analytical techniques to better link chemical and ecological status assessment under the WFD is summarized in figure 2.12.

Figure 2.12: Smart combination of existing approaches for characterizing a water body can support the understanding of connections between chemical contamination and ecological status

Clearly, there are limitations as to what can be reasonably expected from such efforts, with both scientific and practical considerations, such as:

Chemical analysis of freshwaters is limited to what has been looked for, be that through targeted, screening or untargeted analytical strategies. The limitations are specific for each approach;
Complementary use of effect-based methods needs to consider which tests should be used;
Effect-based methods rely on concentrating the dissolved substances in a water sample through solid phase extraction methods. Such methods work well for some organic compounds (non-polar) but not for others (e.g. polar compounds including glyphosate and AMPA) (Reemtsma et al. 2016). Neither metals nor contaminants bound to particles will be detected by the effect-based methods discussed and would thus need separate analysis. This is a significant omission given the relatively widespread failure of metal EQSs (EEA, 2018a; Johnson et al. 2017).

comments (6)

2.8.         Summary

The major advantage of incorporating mixture assessment and biological effect detection is that the effects of chemical pollution can be identified more comprehensively, allowing  further bridging between chemical and ecological status.

Most effects-based methods do not provide conclusive evidence of the chemical(s) responsible. That requires further, site-specific effort, which is where scientific technique bumps into a regulatory approach based on individual substances. Water managers need to, firstly, identify which components of the mixture are the main contributors to the harmful effects, and secondly, to reduce those inputs. However, this approach is not entirely new – “biological oxygen demand” (BOD) has been used many years as an integrated measure of water pollution.

In relation to chemical status assessment under WFD, the inclusion of techniques more sensitive to chemical pollution is likely to make it more difficult to achieve good chemical status. While this situation may reflect expert opinion based on current scientific knowledge as to “real chemical status” it would represent further difficulties in communicating progress under the WFD. One option could be for effects-based methods to be used as part of ecological status assessment.


comments (3)