Reference scenarios for exposure to plant protection products and invertebrate communities in stream mesocosms.

Higher tier aquatic risk assessment for plant protection products (PPPs) is often based on pond-like mesocosm studies in which transient and dynamic PPP exposure scenarios as observed in lotic systems are hardly achievable. Thus, the present study presents dynamic PPP exposure scenarios at different time scales under flow-through conditions as typical for streams in agricultural landscapes. The stream mesocosm setup allows testing the influence of spatial gradients of exposure over the length of themesocosms. The use of the fluorescent tracer uranine revealed the hydraulic processes generally underlying peak- and hour-scale exposure scenarios and demonstrated an optimized application technique to achieve stable day-scale exposures. Furthermore, to account for potential reactions of invertebrates to PPP exposures in streams (e.g. avoidance behavior and drift), the present study thus aimed at a comprehensive evaluation on how PPP exposure and the establishment of invertebrates can be advanced within streammesocosm testing. For both, peak- and hour-scale exposure as well as the experiments considering the establishment of invertebrates, the presented compilation of experiments was able to highlight the influence of aquatic macrophyteswithin streammesocosms. Since the field relevance of the higher tier aquatic risk assessment for PPPs relies qualitatively on the presence of potentially sensitive or vulnerable species, those species were especially considered. Thus, the establishment of aquatic invertebrates in nondosed streams was evaluated with respect to (i) the presence of different aquatic macrophytes and (ii) the duration of the pre-experimental period. The present study highlights the beneficial influence of complex-structured macrophytes and prolonged pre-experimental periods on the abundance of invertebrate taxa. Furthermore, population dynamics were evaluated statistically by simulating PPP-related declines of 30, 50 and 70%. Thereby,