Are safety margins for pesticide mixture risks in German small streams sufficiently large? Results from dynamic model-based analyses of monitoring results

Abstract Pesticide mixtures in the environment are increasingly receiving attention in regulatory environmental risk assessment. The simultaneous presence of fluctuating levels of multiple pesticides can be relevant for the risk assessment of potential toxic effects on aquatic organisms. Although the need to incorporate mixtures in environmental risk assessment has been highlighted for years, procedures are still based on single-substance analysis. We combined dynamic modeling based on the general unified threshold model of survival (GUTS) with results from a three-month monitoring campaign that recorded several pesticides in small German streams. We considered regular and event-driven sampling from 101 sites and 19 pesticides (11 insecticides, 8 fungicides) for the calculation of risk. Measured exposure concentrations were interpolated using three exposure scenarios, filling knowledge gaps about the pesticide time course between samplings. We combined GUTS modeling with the concepts of concentration addition for mixtures of pesticides with similar modes of action (MoAs) and treated mixtures of pesticides with different MoAs according to the independent action concept. Assuming identical damage dynamics per MoA group, models were calibrated on single representative pesticides per group, markedly reducing the calibration effort. The risk to aquatic macroinvertebrates was quantified using the concept of lethal profiles (LP50), which can be interpreted as margins of safety. The results showed that considering pesticide mixtures increased the risk by one to two orders of magnitude relative to single pesticides. Some monitoring sites showed nonnegligible risks, while risk levels varied across exposure scenarios. Particularly, neuroactive pesticides contributed to the mixture toxicity potential. We identified methiocarb, thiacloprid, and epoxiconazole as the key risk drivers within each MoA group. Our study highlights the importance of considering pesticide mixtures for a realistic risk assessment. The retrospective analysis of existing monitoring data could provide a way forward to a more proactive, systems-based approach to prospective environmental risk assessment.