The ecological risk assessment (ERA) of chemicals aims at quantifying the likelihood of adverse ecological effects posed on populations and the communities they comprise (sensu Forbes et al. 2011). Effects caused by the exposure of organisms to toxicants can however to a great extent depend on environmental scenarios as well on the states, behaviors and interactions of organisms with consequences for individual life history, population responses and community dynamics. In the past, these aspects have often been ignored in ecotoxicology and the calculation of adverse effect concentration, such as the predicted no effect concentration (PNEC) or regulatory acceptable concentration (RAC), which classically focuses on toxicant concentration responses while keeping the environmental conditions as constant as possible. Therefore, calls for more ecological realism in ecotoxicology reached the scientific and regulatory community more than a decade ago. As we cannot test the effects of all toxic compounds on all species in all possible environments, we need predictive modeling approaches for ecological risk assessment as recommended by the EU commission (SCHEER 2013).
The prediction of how communities perform under environmental change generally requires an understanding of how macro scale system behavior, such as population dynamics and functional community diversity, emerges from the behavior of its lower scale components. IBMs are thus playing an increasing role in both basic and applied science (DeAngelis and Mooij 2005, Grimm and Railsback 2005, Stillman et al. 2014). A major drawback of IBMs is however that model design differs considerably among species. Analyzing model structure or obtaining general insights into natural systems from IBMs is generally inefficient (Grimm 1999), although standardized protocols for model development and documentation exist (Grimm et al. 2006, Schmolke et al. 2010). For applications in ERA the development of standardized or general model designs has thus been recommended by the European Food Safety Authority (EFSA 2014). Instead of developing models for each species from scratch, standard designs can facilitate the re-usability of IBMs and allow more general insights into natural systems (Berger et al. 2002). It has been argued that standard models will make model development and communication more efficient, and there would be no need for detailed justification every time a model design is re-used because it has been applied before (Martin et al. 2013). Moreover, standard models will facilitate the comparison of different species and ecosystems, as the variability among systems can be ascribed to species-specific traits or system-related factors rather than to any detail of the model structure. Strong moves in the direction of standardizing IBMs have been made based on metabolic theories (Sibly et al. 2013, Martin et al. 2013, Gergs et al. 2014).
- Identification of relevant ecological scenarios as a basis for model parameterization and application to be stored in dedicated data-bases.
- Development and implementation of a standardized individual-based community model, with a basis in dynamic energy budget theory (Kooijman 2010), that is able to represent aquatic communities typical for major habitat typologies in the European Union.
- Explore the trait-based community IBM (see above) and a differential equation based community model (already existing) as well as statistical and process-based effect models to compare community level toxicity model outcomes to conventionally calculated PNECs.
- Develop a conceptual framework for adopting ecological risk assessment using the REACH regulation as an example.