Ecological risk assessment example

The remainder of this section details the potential application of multivariate methods in the selection of endpoints and in the evaluation of exposure and effects of stressors in ecosystems. Particular reference is made to the application of these methods to the current framework for ecological risk assessment. Examples of the use of multivariate methods in detecting effects and in selecting important measurement variables are covered using both field surveys and multispecies toxicity tests. 

Another important reason for using multiple scenarios is to represent major sources of variability, or what-if scenarios to examine alternative assumptions about major uncertainties. This can be less unwieldy than including them in the model. Also, the distribution of outputs for each separate scenario will be narrower than when they are combined, which may aid interpretation and credibility. A special case of this occurs when it is desired to model the consequences of extreme or rare events or situations, for example, earthquakes. An example relevant to pesticides might be exposure of endangered species on migration. This use of multiple scenarios in ecological risk assessment has been termed scenario analysis, and is described in more detail in Ferenc and Foran (2000). 

The extent to which the five criteria are evaluated depends on the scope and complexity of the ecological risk assessment. However, understanding the underlying assumptions and science policy judgments is important even in simple cases. For example, when exceedence of a previously established decision rule such as a benchmark stressor level or water quality criterion is used as evidence of adversity, the reasons why exceedences of the benchmark are considered adverse should be clearly understood. 

The majority of toxicity test data are generated using species from the northern hemisphere (i.e., Holarctic). For example, 9 of the 12 freshwater fish species used in the ecological risk assessment of atrazine (Solomon et al. 1996) and 27 of the 40 freshwater fish species used in the risk assessment of copper (Brix et al. 2001) are from Holarctic habitats. Relatively few data are available for southern hemisphere species, and consequently risk assessments conducted to protect southern hemisphere ecosystems have to utilize toxicity data obtained using northern hemisphere species (Muschal and Warne 2003). Does this matter Based on the limited data currently available, it would appear not. 

GIS may also be used to assess the site-specific bioavailability of toxicants and their ecological risks. For example, Prusha and Clements (2004) related metal concentrations in the lotic insect Arctopsyche grandis to physicochemical characteristics measured in 16 streams. GIS was used to calculate landscape attributes in... 

Any ecological risk assessment at the landscape level has to start with the question, What has to be protected This protection aim preferably needs to include a spatial component (e.g., protecting the aquatic biodiversity from pesticide stress in watercourses neighboring agricultural fields). It may also include a temporal component for example, consider only effects on the densities of aquatic populations to be acceptable in drainage ditches neighboring agricultural fields that show full recovery within a certain time period (e.g., 8 weeks) but do not allow these effects in main watercourses connected to these ditches (see Section 1.3.1 in Chapter 1 for a discussion on this topic). 

Ecological risk assessments cannot be done without applying extrapolation methods. Sufficient data to execute such risk assessments are usually lacking. For example, there may be no toxicity data for the suspect substance, the tested species may differ from the species in the assessed ecosystem, exposure is to single substances in test systems but mixtures occur in the field, or risks are to be assessed for communities rather than for species. The lack of data is a consequence of practical and ethical considerations. 

Ecological risk assessment of chemical mixtures thus has to deal with a variety of field phenomena, a possible range of assessment endpoints, and a variety of assessment approaches. Moreover, there exists a huge variety in the regulatory questions and problem formulations addressed in ecological risk assessment of chemical mixtures. Examples include the protection of specific species against well-defined mixtures (like PCBs and PAHs), the protection of an undefined concept like the ecosystem, and retrospective assessments for highly or diffusely contaminated systems. 

Common Whole Mixtures There are few systematic studies of mixtures that are strictly based on the approach of the mixture of concern or similar mixtures as defined under human risk assessment of mixtures. Most ecological effect studies have more characteristics in common with a component-based or unique whole mixture approach than with the common mixture approach. A rare example of the common whole mixture approach in ecological risk assessment is the hydrocarbon block method. In this case, mixture effects are predicted on the basis of partial characterization of hydrocarbon mixtures. The hydrocarbon block method is used to determine the risks of a total hydrocarbon mixture on the basis of discriminating different chain length fractions of hydrocarbons, for each of which toxicities are known (King et al. 1996). 

There is much dissimilarity between the fields of human and ecological risk assessment, but many of the differences are not typical for mixtures. Examples are differences in assessment endpoints (individuals vs. species or communities), in exposure routes and media (oral, inhalatory, and dermal for humans vs. aquatic or terrestrial for ecosystems), and in the level of mechanistic understanding (generally larger in human than in ecological studies). 

When it comes to mixtures, an important development is the use of the internal dose as a dose metric, particularly in human assessments. The internal dose is either measured directly or modeled using PBPK models, for example, as a blood or a target tissue concentration. Application of an internal dose metric makes it possible to account for 1) interindividual variability in toxicokinetics, 2) temporal variations in exposure patterns, and 3) interactions between substances during absorption, metabolism, and transport. In ecological risk assessment, internal doses are sometimes measured but rarely modeled with PBPK models. The awareness is growing that the internal dose is a useful metric but the use in formal risk assessment procedures is still limited, for separate compounds as well as for mixtures. 

Tier 2 assumes either a uniform MOA for all compounds (i.e., concentration addition) or a complete nonuniform set of modes of action (i.e., response addition). Limited information on the MOA is typically available to use in the assessment, and the techniques are relatively simple. CA in tier 2 differs from that in tier 1 by using the full-dose-response curve. First, the concentration of the components is expressed in comparable units. Subsequently, these units are summed and a dose-response model is applied to predict the response. Examples include the application of RPFs, TEFs, and toxic units. These techniques are commonly used in human as well as in ecological risk assessment of mixtures, though the use of whole curve estimates is by far less common than the use of point estimates (tier 1). 

Tier 3 involves the use of both CA and R A models together (mixed-model approaches). This approach differs from the previous tiers by using detailed information on the modes of action for the different mixture components as well as full-curve-based modeling approaches. Mixed models are used in human as well as ecological assessment. An example of mixed-model approaches in ecological risk assessments is the approach proposed for assemblages (De Zwart and Posthuma 2005) a similar approach has been proposed by Ra et al. (2006) see Chapter 4 and Figure 4.2. 

Tier 4 includes all methods that go beyond CA or RA and attempt to provide some kind of mechanistic explanation for the mixture effects, including potential interactions between the mixture components. It requires detailed information on the toxicokinetic and toxicodynamic processes involved. The diversity of models that belong to this category is huge. Examples from human mixture assessment include the application of PBPK and BRN models. In ecological risk assessment, it may involve the consideration of multiple modes of action per mixture component as well as the assumed characteristics of sets of receptor species. Therefore, tier 4 methods only apply to problems that are defined in a very specific way (regarding site, species, compounds), and where an accurate result is preferred over a conservative one. 

It is evident that some techniques do not have conceptually similar equivalents with various levels of complexity. Hence, tiering is not (yet) possible for all problem definitions. Moreover, it is clear that human risk assessment can sometimes operate on a higher conceptual tier than ecological risk assessment, for example, when BRN modeling and PBPK models are used. On the other hand, ecological risk assessment approaches may be sometimes more diverse, and can be better tailored to a risk assessment problem and its context. 

In the previous sections, different methods for human and ecological risk assessment of mixtures have been reviewed and discussed. Some typical mixture issues were briefly mentioned within the context of this methodological review, but they have not been extensively discussed. Examples of such issues are 1) exposure assessment, 2) sufficient similarity, 3) interactions vs. additivity, 4) QSARs, 5) uncertainties, and 6) risk perception of mixtures. These topics are discussed in more detail in the following sections. 

There are many concepts in use for the assessment of risks or impacts of chemical mixtures, for both human and ecological risk assessment. Many of these concepts are identical or similar, for example, whole mixture tests, (partial) mixture characterization, mixture fractionation, and the concepts of CA and RA. 

There are many concepts in use for the assessment of risks or impacts of chemical mixtures, both for human and ecological risk assessment. Many of these concepts are identical or similar in both disciplines, for example, whole mixture tests, (partial) mixture characterization, mixture fractionation, and the concepts of CA and RA (or I A). The regulatory application and implementation of bioassays for uncharacterized whole mixtures is typical for the field of ecological risk assessment. The human field is leading in the development and application of process-based mixture models such as PBTK and BRN models and qualitative binary weight-of-evidence (BINWOE) methods. Mixture assessment methods from human and ecological problem definition contexts should be further compared, and the comparison results should be used to improve methods. 

The main focus of the ecological risk assessment is to minimize undesired events caused by chemicals. Species sensitivity distribution (SSD) is an example of an ecotoxicological method which is based on such events at above the no-effect level/concentration. We can assume that within a community species differ in... 

It is important to clearly describe and quantitatively estimate the assumptions and uncertainties involved in the evaluation, where possible. Examples include natural variability in ecological characteristics and responses and uncertainties in the test system and extrapolations. The description and analysis of uncertainty in characterization of ecological effects are combined with uncertainty analyses for the other ecological risk assessment elements during risk characterization. 

---