Extrapolation of Laboratory-Measured Effects to Fish Populations in the Field

Abstract

The aim of prospective environmental risk assessments of chemicals is to protect the environment from unwanted negative effects following the use of a given chemical product. The protection goal of relevant chemical regulations is primarily at the population level; however most toxicity data is often only available at the individual level from laboratory studies. Population modelling is one method used to link the effects of chemical exposure observed in the laboratory on individuals, to predicted effects on natural populations and the relevant protection goals. This chapter presents a method to develop a fish population model and is based on the following steps as outlined in the “modelling cycle”: (a) define the purpose of the model, (b) conceptualize the model, (c) formalize the model, (d) implement the model, (e) verify the model, (f) calibrate the model, (g) analyze the model (including sensitivity, uncertainty, and validation) and (h) communicate the model to all interested stakeholders (e.g., scientific community, regulatory authorities and the public). Though much of this chapter is based on the use of population models in chemical risk assessment, population models may be developed for a wide range of purposes (e.g., conservation of endangered species or fisheries stock assessments). While it is impractical to cover all these variations within this chapter, the general strategies employed to develop models within the modelling cycle are broadly applicable and examples are provided where necessary to illustrate each point. This chapter focuses on the development of individual-based population models (IBMs), though other model types (e.g., matrix models, TKTD models) are discussed and much of what is presented in this chapter for IBMs is equally applicable to these other modelling options.

You should use this chapter to develop an individual-based fish population model, though it should be noted that the nuances related to each model due to differences in research question between model developers will lead to models with different structure and outputs. Population modelling overcomes many of the challenges that may beset empirical studies performed to answer the same question at the same spatial and temporal scale (e.g., cost, time, interpretation of results) and as such is a powerful tool in further developing our ecological understanding. Adherence to good modelling practice when developing future models (such as using the approach outlined in this chapter) will ensure that population modelling becomes more widely accepted in the scientific and regulatory communities and population modelling may take its rightful place as a unique tool providing novel insights into population ecology and their application in the risk assessment of chemicals.