Mechanistic models to predict pesticide stress on Daphnia magna populations – an intermediate tier tool for ecological risk assessment

For pesticides, complex and costly higher tier tests are required if risks are identified in the lower tier tests. These higher tier tests can possibly be avoided by the application of mechanistic models e.g. Toxicokinetic-toxicodynamic (TKTD) models or population models such as individual-based models (IBM) based on dynamic energy budget (DEB) theory. These models follow solid scientific knowledge to extrapolate effects observed in laboratory conditions to effects on real populations. TKTD models in particular can be used to predict (sub)lethal effects on the modelled organism, as acknowledged in a recent Scientific Opinion by EFSA. A major advantage of TKTD models is that they can be used to predict toxic effects under time-varying exposure conditions, making them an ideal tool for pesticide risk assessment. As a case study for such a mechanistic approach, DEB-IBMs and TKTD models were used to predict the population dynamics of Daphnia magna population under pesticide stress. Two pesticides were chosen: a copper-based pesticide and endosulfan. TKTD models were employed to predict the pesticide effects observed in toxicity tests: the General Unified Threshold model for Survival (GUTS) was used for lethal effects and DEBtox for sub-lethal effects. In the case of DEBtox, pesticide stress is associated with a certain physiological mode of action (PMoA), leading to effects on sub-lethal endpoints. For copper, increased costs of growth was identified as the most likely PMoA for effects on growth and reproduction. For endosulfan, a reduced reproduction efficiency explained the observed effects best. GUTS and DEBtox were included in the DEB-IBM population model to predict pesticide effects on D. magna populations. Because pesticides have a very specific use and emission pattern, FOCUS scenarios were used to simulate population dynamics for realistic exposure scenarios. For these scenarios, effect concentration were determined for population endpoints (e.g. maximum population growth rate, carrying capacity) and compared with individual-level endpoints. Mechanistic models such as DEB-IBMs and TKTD models allow to extract more ecologically relevant information from lower tier tests and can thus help inform risk assessors on the ecological risk of a pesticide before performing higher tier tests.