Towards the development of standard models for ecological risk assessment: an agent-based model of fathead minnow

• We present a dynamic-energy-budget and agent-based model of fathead minnow. • Our model could be used as a standardized tool for fish higher-tier risk assessment. • We estimated population-level effects of exposure to a fungicide. • Dynamics are defined by the interplay between exposure and density dependence. • Similar life-history traits lead to similar population-level responses. Population models are recognized as an essential tool for higher-tier risk assessment, yet they are underutilized in North America. This arises, in part, from the lack of standardized models with which users and risk assessors can familiarize themselves, coinciding with an absence of standardized protection goals for populations. In this paper, we present the adaptation of a previously published population model for aquatic risk assessment to the species fathead minnow (FHM). We argue that this model can support a standardized modeling workflow for higher-tier risk assessment in freshwater fish. The model represents the life cycle of FHM using the Dynamic Energy Budget theory and FHM population dynamics. It can reproduce laboratory experiments and population dynamics in realistic environments via multiple modules accounting for different effect pathways. Moreover, it is sufficiently flexible to be adapted from representing FHM to other species of interest. We applied our model to estimate population-level effects of exposure to the fungicide chlorothalonil. The exposure profile was constructed based on realistic hydroxychlorothalonil exposure levels, which is the major metabolite of chlorothalonil. We show that the population dynamics are defined by the interplay between exposure and density-dependent compensation effects. Comparing our findings with previous results, we found that similar Cyprinid life-history traits lead to similar population-level responses to chemical exposure, potentially expanding the use of the model to data-poor species. Finally, we calculated three population-level outputs and suggested them as candidates for successful endpoints across different risk assessment studies.