Development and Application of Modelling Approaches for Realistic Assessments on Population Impacts of Endocrine Disruption in Fish.

Chemical exposures threaten the health of freshwater ecosystems worldwide. In particular, endocrine disrupting chemicals (EDCs) are of concern because of their ability to cause sub-lethal effects on organisms at low, including environmentally relevant, concentrations. The susceptibility of fish populations to the effects of these chemicals depends on exposure risk, physiological susceptibility, and population resilience. Population models can explicitly incorporate these factors into environmental risk assessments (ERAs) to improve realism and identify potentially vulnerable species. In this thesis, modelling tools were developed and evaluated for the three-spined stickleback (Gasterosteus aculeatus) to advance understanding of the ecological relevance of EDC effects.The thesis begins with a critical review of the current status of EDCs in freshwater ecosystems worldwide and their effects on individual fish and their populations. The potential for different modelling techniques to provide realistic ecological assessments for EDCs is then explored. Individual-based models (IBMs) are used throughout this thesis to provide case study specific chemical assessments. Reproductive endpoints, including disruption of breeding behaviours, are used to extrapolate EDC effects from individuals to the population level. Findings included the importance of considering behavioural endpoints within chemical assessments, since disruption of breeding behaviours caused significant reductions in population abundance. Moreover, it was identified that the breeding strategy of the stickleback makes it particularly vulnerable to chemicals which directly affect reproductive output. The chemical exposure regime and density dependent processes determined whether the population recovered post-exposure. Empirical experimental exposures were used to investigate the interactive effects of EDC exposure and food limitation on somatic growth in early life stages. An energy budget model was then developed and used to further explore the mechanisms underlying this observed effect. The combined empirical and modelling results suggested that fish can adapt their physiology (by reducing physical activity) to cope with the effects of multiple stressors. Finally, in order to explicitly incorporate environmental conditions into population level assessments, a model was developed combining the energy budget model with the stickleback IBM. This model allows analyses on direct effects of the EDC as well as the additional metabolic effects associated with the chemical exposure. The effects of two case study EDCs (an oestrogen and an androgen) on individual fecundity were simulated in low and high food availability environments in order to explore how environmental conditions affect population susceptibility. The findings illustrate that the underlying mechanism of the EDC effect and environmental conditions can affect the susceptibility of populations to EDC exposures. This thesis has developed novel models for use by both researchers and risk assessors for application in realistic population level assessments of chemical risks. The findings presented have important implications for understanding the ecological relevance of EDC exposures for fish populations