Probabilistic risk assessment of insecticide concentrations in agricultural surface waters: a critical appraisal.

Due to the specific modes of action and application patterns of agricultural insecticides, the insecticide exposure of agricultural surface waters is characterized by infrequent and short-term insecticide concentration peaks of high ecotoxicological relevance with implications for both monitoring and risk assessment. Here, we apply several fixed-interval strategies and an event-based sampling strategy to two generalized and two realistic insecticide exposure patterns for typical agricultural streams derived from FOCUS exposure modeling using Monte Carlo simulations. Sampling based on regular intervals was found to be inadequate for the detection of transient insecticide concentrations, whereas event-triggered sampling successfully detected all exposure incidences at substantially lower analytical costs. Our study proves that probabilistic risk assessment (PRA) concepts in their present forms are not appropriate for a thorough evaluation of insecticide exposure. Despite claims that the PRA approach uses all available data to assess exposure and enhances risk assessment realism, we demonstrate that this concept is severely biased by the amount of insecticide concentrations below detection limits and therefore by the sampling designs. Moreover, actual insecticide exposure is of almost no relevance for PRA threshold level exceedance frequencies and consequential risk assessment outcomes. Therefore, we propose a concept that features a field-relevant ecological risk analysis of agricultural insecticide surface water exposure. Our study quantifies for the first time the environmental and economic consequences of inappropriate monitoring and risk assessment concepts used for the evaluation of short-term peak surface water pollutants such as insecticides.