How to Select a Model for a Risk Assessment

Depending on their complexity and ease of use, some ecological models are more appropriate for screening-level ecological risk assessments, whereas others are best reserved for detailed assessments. For example, simple models of population dynamics are most appropriate for screening-level ecological risk assessments. Ecosystem and landscape models are generally reserved for detailed ecological risk assessments because of the greater effort and expense involved in applying these models.

Ecotoxicological model selection is based on the objectives of the ecological risk assessment and the data available for modeling. The following factors should be considered when choosing an ecotoxicological model to address a specific risk issue:

• Kind of question. Is the ecological risk assessment assessing impacts of chemical contamination already present at a site or potential risks from uses of new chemicals? Is the ecological risk assessment addressing chemical releases and contamination, or cleanup/restoration issues?

• Level of biological organization. Does the assessment endpoint deal with populations, communities, ecosystems, or landscapes?

• Specific endpoints and ecological relevance. Are ecological structures (e.g., habitat pattern, species abundances, or distributions) or processes (e.g., ecosystem productivity, population recruitment) of interest? How related are the model results to the endpoints that are used in ecological

Single species

Spatial structure

Individual structure

Age-stage structure

Single species

Aggregated

Single pooled group

Spatially explicit

Multiple pooled groups

Individuals tracked

Single pooled group

Multiple pooled groups

Pooled

Discrete

Pooled

Discrete

Age-stage structure

Pooled

Discrete

Scalar Life-history metapopulation metapopulation model model

Density effects DI

Scalar Life-history metapopulation metapopulation model model

Density effects DI

Individuals tracked

Legend

Major model categories DI Density independent DD Density dependent

Figure 2 Classification tree for single-series models. Copyright (© 2002) Ecological Modeling in Risk Assessment: Chemical Effects on Populations, Ecosystems, and Landscapes, by Pastorok RA, Bartell SM, Ferson S, and Ginzburg LR (eds.). Reproduced by permission of Routledge/Taylor & Francis Group, LLC.

Multiple species

Major components

Spatial structure

Habitat Legend

Biotic

Aggregated

Aggregated

General

Major model categories

Abiotic and biotic

Aggregated

Spatially explicit

Aggregated

Spatially explicit

Aquatic Terrestrial General

Aquatic Terrestrial General

Aquatic Terrestrial General

Figure 3 Classification tree for multispecies models. Coptright (© 2002) Ecological Modeling in Risk Assessment: Chemical Effects on Populations, Ecosystems, and Landscapes, by Pastorok RA, Bartell SM, Ferson S, and Ginzburg LR (eds.). Reproduced by permission of Routledge/Taylor & Francis Group, LLC.

risk assessments.? If the model results are not directly relevant endpoints, can the endpoints be easily calculated from model results.

• Screening versus detailed. Is the ecological risk assessment a screening-level assessment or a more detailed assessment?

• Level of realism. Does the model incorporate key processes known to be important in the system it simulates, as well as key factors that affect each of these resources? Are the assumptions realistic with respect to the ecology of the system? Is it necessary to model specific mechanistic processes? Should exposure-response relationships be explicitly incorporated into the model?

• Spatially explicit or aggregated. Is the spatial distribution of ecological entities of interest, so that a spatially explicit model is needed?

• Individual organism or groups. Is the assessment population a threatened or endangered species? Should the model simulate the distribution, behavior, and characteristics of individual organisms, or should it simulate groups of organisms (i.e., local populations or communities)?

Additional criteria for evaluating the suitability of a model for a specific problem include the following practical considerations:

• Flexibility. Can the model accept alternative formulations? For example, can the model be applied to species and systems other than those for which it was originally developed? Can different life histories, distributions, or habitat types be easily modeled?

• Treatment of uncertainty. Does the model incorporate uncertainty? How easy is it to parse effects of natural variability and measurement uncertainties?

• Degree of development and consistency. How easy is it to understand what the model does, so that the underlying mechanisms can be checked for errors and bugs? Has the model been tested or validated?

• Ease ofparameter estimation. How easy is it to estimate the required model parameters given the type of data available? Are accepted sampling or statistical methods for estimating model parameters included?

• Regulatory acceptance. How likely are regulatory agencies to accept results of an analysis with the model? Has the model been used by any regulatory agencies?

• Credibility. Does the model have scientific and technical credibility? Is it widely known? Has it been used in support of published research?

• Resource efficiency. How much time and effort would be needed to apply the model in a particular case?

Solar Power

Solar Power

Start Saving On Your Electricity Bills Using The Power of the Sun And Other Natural Resources!

Get My Free Ebook


Post a comment