Assessment endpoint

According to USEPA, the key aspect of the ERA is the problem formulation phase. This phase is characterized by USEPA as the identification of ecosystem components at risk and specification of the endpoints used to assess and measure that risk [13]. Assessment endpoints are an expression of the valued resources to be considered in an ERA, whereas measurement endpoints are the actual measures of data used to evaluate the assessment endpoint. 

The US Environmental Protection Agency (USEPA 1998) describes problem formulation as an iterative process with 4 main components integration of available information, definition of assessment endpoints, definition of conceptual model, and development of an analysis plan. These 4 components apply also to probabilistic assessments. In addition, it is useful to emphasize the importance of a 5th component dehnition of the assessment scenarios. The relationships between all 5 components are depicted in Figure 2.1. Note that the bidirectional arrows represent the interdependency of the different components and imply that they may need to be revised iteratively as the formulation of the problem is rehned. 

Assessment endpoints are measurable ecosystem characteristics that represent management goals (USEPA 1998). They should define... 

The assessment endpoint should be not only measurable (at least potentially) but also modelable. Defining a modelable endpoint is likely to require close discussion between an assessor (who knows what they can model) and a risk manager (who knows what they want to protect). Sometimes the assessment endpoint is only indirectly related to the management goal, for example, if the assessment endpoint is a risk to individuals, but the aim is to protect population sustainability. In such cases, qualitative inference will be required to interpret the assessment result. This inference will need to be done jointly by the risk assessor and risk manager. It is likely to involve substantial uncertainty, which will have to be taken into account qualitatively when producing a narrative description of the assessment outcome. This step should be identified as part of the conceptual model. 

If the assessment is to be probabilistic, the risk assessor and risk manager should consider together how this influences the definition of the assessment endpoint. Suter (1998) suggests 5 questions for the risk assessor to ask the risk manager to help define the assessment endpoint ... 

Should any assessment endpoints be expressed as probabilities Suter (1998) points out that it can be confusing to use the term probability in defining assessment endpoints because it is unclear whether it relates to variability or uncertainty, so it will be helpful to distinguish these in the discussion with the risk manager. 

Questions 2 and 3 imply a choice between expressing effects in terms of magnitude, frequency, and certainty. In practice, the assessment endpoint may often need to be dehned in terms of 2 or 3 of these dimensions. For example, it may be desirable to estimate the proportion of species (frequency) that will experience different levels of mortality (magnitude), and to provide confidence limits (certainty). Indeed, the risk manager s questions may imply an assessment endpoint with more than 3 dimensions, for example, if it is desired to express frequency in terms of space (e.g., number of hectares) and time (proportion of years). The dimensionality of the assessment endpoint will have major implications for all aspects of the analysis and for communication of results, so it is essential to discuss it carefully with the risk manager at the outset to ensure it meets their needs. 

If the assessment endpoint is a distribution, or a statistic from a distribution (e.g., 95th percentile), it is essential to be clear how the distribution is interpreted (Suter 1998, p 129). If it is a frequency distribution, to what statistical population does the distribution refer For example, does the distribution represent a population of individuals, an assemblage of species, a number of locations treated with pesticides, or a series of time periods The answer to this question has substantial implications for the structure of the assessment model and the types of data required. 

It is essential to define the assessment scenario within which the assessment endpoint will be assessed. The assessment scenario should specify the spatial, temporal, and ecological boundaries within which the endpoint is assessed, since these have substantial implications for the structure of the assessment model and the scope of the input data. The assessment scenario should also describe those aspects of the ecosystem that are relevant to the assessment, that is, those aspects that have an influence on the mechanisms of exposure and effects that will be assessed. This step is important in all ecological risk assessments it places the assessment activity into the real context of an ecosystem, helps to prevent construction of inappropriate models, and helps with interpretation and communication of results. 

The choice of assessment scenario, like the assessment endpoint, is likely to be implied by the management goal and should be made in close consultation with the risk manager, to ensure it meets their needs. 

For pesticide risk assessments, it may often be necessary to assess impacts of the same pesticide used in different crops, in different seasons, in different geographic regions, and on different species and ecosystems. This will require the use of multiple scenarios and possibly multiple assessment endpoints. 

A set of risk hypotheses that describe predicted relationships among stressor, exposure, and assessment endpoint response, along with the rationale for their selection... 

It may also be useful to include in the conceptual model other lines of evidence that are relevant to the assessment endpoint. This may help to highlight the contribution that other lines of evidence can make and promote more effective gathering and use of such information. 

Defining an appropriate relationship between the units of analysis and the assessment endpoint... 

Finally, it is essential that the units of the analysis are suitable for generating the assessment endpoint. 

Tabulating the temporal, spatial, and biological scales of each component of the assessment may help to identify appropriate units of analysis, show how they relate to real-world processes, and check their compatibility with the assessment endpoint and hence the management goal (e.g., Table 2.1). 

Assessment endpoint Local populations of single focal species, chosen to represent a range of similar species around cornfields Likelihood, frequency, and magnitude of bird kills caused by pesticide exposure... 

Defining the Relation between Units of Analysis and the Assessment Endpoint... 

Careful construction of the conceptual model diagram, and the use of a tabular approach such as Table 2.1, should help to avoid these problems. The diagram should show clearly the point at which individual exposure is used to predict individual effects and the process by which individual effects are aggregated to generate the risk estimate. In addition, it should be remembered that the risk estimate may be combined quantitatively or qualitatively with other lines of evidence to address the assessment endpoint. 

Systematically examine each risk hypothesis for each type of analytical uncertainty, include them in the description of the risk hypothesis, and state in which direction they are expected to affect the assessment endpoint. 

Before deciding to treat a line of evidence separately, consideration shonld be given to whether it can in fact be directly incorporated into the quantitative assessment. For example, it may be possible to use Bayesian updating to incorporate information from field studies or monitoring if they provide direct measurements of the assessment endpoint, or of the intermediate model. 

USEPA] US Environmental Protection Agency. 2003. Generic ecological assessment endpoints (GEAEs) for ecological risk assessment. Risk Assessment Forum. Washington (DC) USEPA. 

Outlining protection goals including resources to be protected and assessment endpoints... 

Essential prereqnisites for a probabilistic risk assessment are a well thought-out problem formulation and a clear definition of the assessment endpoints. The probabilistic approach according to its very natnre aims at making predictions on quantities or the occurrence of certain events. Snch qnantities and events must be specified precisely such that, at least in principle, there is no doubt on what the quantity is or whether the event happened (Morgan and Henrion 1990). 

Usually models are created for a certain purpose, and that purpose drives their structure, level of detail, level of complexity, etc. A model may be excellent, but it must not be used for inappropriate purposes. If the output of the model does not match the assessment endpoint and the questions raised in the problem formulation phase, then the model obviously is not suitable for the specihc case. 

Formulating the assessment problem well is an essential foundation for risk assessment. The workshop considered how the use of probabilistic models and uncertainty analysis affects problem formulation and its main components the integration of available information, definition of the assessment endpoint, specification of the conceptual model, and planning of the analysis phase. 

Define the assessment endpoint precisely, in terms of probabilities, e.g., the probability of a given level of mortality in the exposed population. 

Ensure that the assessment endpoint is capable of being modeled and has attributes that are measurable. 

Ideally, define the assessment endpoint so that it relates directly to the management goal. If this is not practical, (e.g., if the management goal refers to population sustainability but the assessment endpoint refers to individual mortality), define in advance how the assessment endpoint will be interpreted. If this involves subjective judgments then consider the use of formal methods. 

Regardless of the method used, the basis of the final risk characterization must be explicit. All components and sources of evidence should be described. The explicit linkage between the analysis results and the assessment endpoints must be clearly but adequately stated. Tandem presentation of conventional methods (e.g., ad hoc weight of evidence) and formal methods (e.g., Bayesian, meta-anal-ysis) are recommended to enhance understanding. This is intended to facilitate acceptance of unfamiliar approaches, not to imply that the conventional methods are a touchstone. 

Make sure that the problem is thoroughly formulated and the assessment endpoints precisely stated. 

Assessment endpoint An explicit expression of the enviromnental value that is to be protected, operationally defined by an ecological entity and its attributes. For example, sahnon are valued ecological entities reproduction and age-class structure are some of their important attributes. Together salmon reproduction and age-class structure form assessment endpoints. 

Measurement endpoint A measurable ecological characteristic that is related to the valued characteristic chosen as the assessment endpoint. Also known as measure of effect. ... 

---