SEARCH

SEARCH BY CITATION

Population models are considered as one method for chemical risk assessment, especially for addressing protection goals within the ecosystem services concept (EFSA 2010). Such models have a long history in ecology and they are increasingly being used also for the risk assessment of pesticides (Barnthouse et al. 2007; Forbes et al. 2009). However, it is frequently debated to what extent population models are useful for risk assessment and if their output can be considered realistic, or if they produce rather fictitious predictions. Population models are necessarily more complex than standard methods used in pesticide risk assessment. Although the higher complexity is due to the aim to produce a population model that is as realistic as possible, it is understandable that regulators require a detailed evaluation of the realism and suitability of these models. We therefore address the question of how to evaluate whether a model is sufficiently realistic for use in pesticide risk assessment. We believe that a systematic approach may help regulators to evaluate population models.

Make a Full Assessment of Knowledge About Your Species

  1. Top of page
  2. Make a Full Assessment of Knowledge About Your Species
  3. Identify What Is Relevant
  4. Design Validation Experiments to Match Your Data
  5. Address Uncertainties
  6. Conclusions
  7. References

A thorough assessment of the biology of a species is the first step to understand which mechanisms and parameters have to be included in a population model (Wang and Luttik 2012). This will help to identify which information is really reliable and important, and hence, should be included in a model, and what is not well understood or supported by the literature. This is different from standard risk assessments, which are usually based on a relatively small set of data, whereas population models make use of all knowledge in a much more comprehensive way. However, a thorough review is also crucial for identifying all possible mechanisms or data that can be used in a subsequent validation analysis.

Identify What Is Relevant

  1. Top of page
  2. Make a Full Assessment of Knowledge About Your Species
  3. Identify What Is Relevant
  4. Design Validation Experiments to Match Your Data
  5. Address Uncertainties
  6. Conclusions
  7. References

For showing if a model is sufficiently realistic, one first has to identify which factors of the biology of a species are really relevant for the application of the model. In pesticide risk assessment, effects on population abundance and recovery are the main focus. Hence all factors that may affect either of them should ideally be addressed. In all species, survival and reproduction will have a crucial impact on both population dynamics and recovery. Which additional factors are relevant may depend on the species. For example, for assessing the population dynamics in mammals or birds, spatial distribution and behavior may play an important role, whereas for aquatic organisms the spatial distribution may not be as important.

Design Validation Experiments to Match Your Data

  1. Top of page
  2. Make a Full Assessment of Knowledge About Your Species
  3. Identify What Is Relevant
  4. Design Validation Experiments to Match Your Data
  5. Address Uncertainties
  6. Conclusions
  7. References

Usually field or laboratory data from the literature are limited and often the answer to a specific question raised for validation cannot be answered directly by the results from a study. Hence validation experiments have to be designed to match the available data in the literature. For the evaluation of age structure, for example, the age of animals is usually hard to determine in the field. However, age structure can also be analyzed based on capture-mark-recapture experiments, which reveal a minimum age for different age classes. When reproducing the same age classes in a model, such data can be used to validate age structure.

Address Uncertainties

  1. Top of page
  2. Make a Full Assessment of Knowledge About Your Species
  3. Identify What Is Relevant
  4. Design Validation Experiments to Match Your Data
  5. Address Uncertainties
  6. Conclusions
  7. References

Although uncertainty is not addressed in most standard risk assessments (safety factors are often used, but it is unclear how “safe” they are), they are usually a requirement for higher tier risk assessments (EFSA 2009). Figure 1 shows that there are two ways of addressing uncertainty in a model:

  • Uncertainty of input parameters is extrapolated to the model's output. This is the approach frequently used in probabilistic risk assessments: the uncertainty of many parameters (e.g., toxicity, exposure) is combined to predict the uncertainty of the effect predicted by the model. This method is the only possibility if no data are available for confirming the output of a model.

  • Uncertainty is evaluated by validation. If there are data from field or laboratory studies, with which a model can be tested, then such tests can show if a model produces a realistic prediction. Although the model output can be exactly measured (including the variability of parameters), the field or laboratory data with which the model results are compared are uncertain. Hence this uncertainty has to be addressed.

thumbnail image

Figure 1. Two methods to address uncertainty. If reference data for validation are not available, then uncertainties have to be extrapolated to model predictions. If data for validation are available, then the model output can be compared to reference data from the field or laboratory.

Download figure to PowerPoint

Conclusions

  1. Top of page
  2. Make a Full Assessment of Knowledge About Your Species
  3. Identify What Is Relevant
  4. Design Validation Experiments to Match Your Data
  5. Address Uncertainties
  6. Conclusions
  7. References

A systematic approach to evaluate whether a population model is sufficiently realistic may help regulators to accept population models as a standard method in chemical risk assessment. This approach may start with a detailed review of what we know about a species and which are the mechanisms driving its population ecology. In an ideal case, the resulting model could include most of our knowledge about a species. This is considerably different from standard risk assessments, which are usually based on a relatively small number of studies. A systematic analysis of relevant parameters and mechanisms and their uncertainty and variability helps to decide if a model includes all relevant mechanisms and parameters and if they are realistically reproduced by the model. In contrast, standard risk assessments (or environmental fate models) generally include safety factors or worst-case assumptions but no systematic evaluation of variability or uncertainty. Hence, the evaluation of a population model can produce a much clearer picture of the predicted risk and the involved uncertainty than standard risk assessments.

References

  1. Top of page
  2. Make a Full Assessment of Knowledge About Your Species
  3. Identify What Is Relevant
  4. Design Validation Experiments to Match Your Data
  5. Address Uncertainties
  6. Conclusions
  7. References
  • Barnthouse LW, Munns WR, Sorensen MT. 2007. Population-level risk assessment. Boca Raton (FL): Taylor & Francis. p 376.
  • [EFSA] European Food Safety Authority. 2009. Guidance document on risk assessment for birds and mammals. EFSA Journal 7: 14381577.
  • [EFSA] European Food Safety Authority. 2010. Scientific opinion on the development of specific protection goal options for environmental risk assessment of pesticides, in particular in relation to the revision of the Guidance Documents on Aquatic and Terrestrial Ecotoxicology (SANCO/3268/2001 and SANCO/10329/2002). EFSA Journal 8: 18211876.
  • Forbes VS, Hommen U, Thorbek P, Heimbach F, Van den Brink PJ, Wogram J, Thulke H-H, Grimm V. 2009. Ecological models in support of regulatory risk assessments of pesticides: developing a strategy for the future. Integr Environ Assess Manag 5: 167172.
  • Wang M, Luttik R. 2012. Population level risk assessment: practical considerations for evaluation of population models from a risk assessor's perspective. Environ Sci Europe 24: 3.