point. Thus, the criterion mimicks the visual assessment process, taking into account the presence and size of systematic deviations, and whether the model adequately predicts the last data point, as a measure for extrapolation capacity. We find that SFO fits with SWARC < 40 can be considered clearly acceptable; for higher SWARC values, SFO may still be acceptable (particularly if SWARC < 65), but DFOP should also be assessed. Testing of the criterion for metabolite fits showed that, in spite of a weaker correlation to consensus score, the criterion can also be useful for metabolites. Taken together, we provide a novel tool that quantifies the visual assessment of SFO fits. This can guide decision making and thus help to reduce subjectivity in regulatory assessments.
MO120
"Southside"- Bridging the hemispheres - Global use of field trials based on ecoregion similarities between New Zealand, Chile and Europe
B. Gottesbueren, BASF SE / Crop Protection, Environmental Fate Modelling; H. Bayer, BASF SE; K. Platz, BASF SE Agrarzentrum Limburgerhof / Environmental Fate Modeling; B. Erzgraeber, BASF SE; F.P. Donaldson, BASF Corporation / APD/EFR; J. Goulet-Fortin, BASF SA; F. Kröger, Eurofins Agroscience Services GmbH
In European regulations degradation rates in soil (DegT50) from terrestial field dissipation studies TFD studies considered for exposure modeling may originate from "any" sites with soil and climatic conditions similar to Europe. An OECD Ecoregion similarity model (ENASGIPS) had been developed (OECD 2016) for gaining acceptance of field studies conducted in North America to Europe and vice versa, and there is no obvious scientific reason why this cannot be extended to other regions of the world. An experimental and GIS/modeling feasibility study (“Southside”) was initiated to demonstrate if TFD studies conducted in the Southern hemisphere (i.e. New Zealand, Chile) under climatic, soil and cropping conditions similar to conditions in the Northern hemisphere may deliver similar soil degradation rates and DegT50 endpoints than those from Europe. Similar similarity zones were identified between the New Zealand and Chilean sites and EU / NAFTA using the OECD ENASGIPS tool as well as an adapted GIS crosswalk with JRC-EFSA climate and soil maps for EU. The trial sites had soil types ranging from loamy sands, sandy loam, loam and silty loams. In New Zealand the sites were located on the Northern Island having an average annual air temperature of ~ 12-13 °C and an average cumulative annual rainfall of ~ 780- 970 mm. In Chile the sites were located in the Región del Bío-Bío east of Concepción having an average annual air temperature of ~ 14 °C and an average cumulative annual rainfall of ~ 800-900 mm. The terrestrial field dissipation (TFD trials) were conducted according to OECD 232 (DegT50 module, soil covered with sand) with different pesticides at 3 sites in New Zealand and in Chile, having no historic use of these pesticides. All pesticides were applied in commercial formulations as a tank mix together in the same spraying on the same field plots at the same time The field DegT50 were normalised to reference conditions (20°C, moisture pF2) during kinetic analysis according to FOCUS, considering local soil conditions and weather data to estimate soil temperature and soil moisture with the PEARL model (as had been done with the EU studies). The quality indicator values of curve fit to data (Chi2-error) were found to be acceptable. The normalized SFO DegT50 in the “Southside” trials in New Zealand were found to be in the range of those from TFD studies in Europe using the same study design.
MO121
Residues of currently used pesticides in Central Europe arable soils: status quo, reasons and consequences
J. Hofman, Masaryk University, RECETOX / Faculty of Science, RECETOX; P. Kosubova, Central Institute for Supervising and Testing in Agriculture; S. Polakova, Central Institute for Supervising and Testing in Agriculture / Official control section; M. Hvezdova, Masaryk University / Research Centre for Toxic Compounds in the Environment (RECETOX); L. Brodský, Charles University in Prague; L. Bielská, RECETOX, Faculty of Science, Masaryk University / Faculty of Science, RECETOX; K. Brandstätter-Scherr, University of Natural Resources and Life Sciences / Institute for Environmental Biotechnology, Department for Agrobiotechnology (IFA-Tulln); P. Dinisová, AQUATEST Inc.; Z. Simek, L. Skulcova, Masaryk University / Research Centre for Toxic Compounds in the Environment (RECETOX); M. Šudoma, Masaryk University; M. Sáňka, Masaryk University / Central Institute for Supervising and Testing in Agriculture; M. Svobodová, L. Krkošková, J. Vasickova, Masaryk University / Research Centre for Toxic Compounds in the Environment (RECETOX); N. Neuwirthová, Masaryk University
Current agricultural management is usually based on high consumption of pesticides which may bring a lot of environmental problems. Alarming results from monitoring pesticide residues in EU groundwater and surface water evoke the question of whether the arable soil can contain significant contamination as a result of the intensive use of pesticides in the present or past. Therefore, in 2014 - 2017, agricultural soil was monitored at more than 100 locations in the Czech Republic for more than 50 representatives of currently used pesticides, their selected transformation products and also banned atrazine and simazine with their transformation products. The results showed that the contamination of the monitored soils with the analysed pesticides is quite extensive. At least one pesticide was detected in 99% soils and in 81% soils the concentration of at least
one pesticide exceeded the threshold of 0.01 mg/kg. The soils also frequently contained multiple residues: 85% soils contained 3 or more pesticides and 51% soils 5 or more pesticides. Over half the soils (53%) contained at least 2 pesticides exceeding 0.01 mg/kg. The most frequent compounds were triazine herbicides (present in 89% soils), which were also in significant concentrations (47% soils with triazine sum exceeding 0.01 mg/kg). Based on the association with the occurrence of terbuthylazine and crops, it was confirmed that banned toxic simazine is still introduced significantly to the soils as an allowed impurity of massively applied terbuthylazine. Persistent atrazine residues are still a legacy of the past, even over 10 years after its last use. The second most frequent compounds were conazole fungicides (present in 74% soils; 53% soils with conazole sum exceeding 0.01 mg/kg). Although no health or environmental risk analysis has yet been carried out on the data, the results draw attention for potential impacts, because: (a) foreign limits based on risk calculations have often been exceeded; (b) many of these substances are suspected carcinogens or endocrine disruptors; (c) substances occur in mixtures whose (eco)toxicity may be additive or even synergistic. The research was carried out with the support of the GACR (project 15-20065S).
MO122
Does the TOXSWA model simulate reliable concentrations in FOCUS surface water scenarios for a single segment water layer ?
P.I. Adriaanse, Alterra Wageningen University and Research Centre; W. Beltman, Alterra Wageningen UR
For the aquatic risk assessment of pesticides according to EU Regulation 1107/2009 exposure concentrations are calculated in ponds, ditches and streams in ten FOCUS surface water scenarios distributed across the EU
(https://esdac.jrc.ec.europa.eu/projects/surface-water). Currently, these scenarios are based on simulation periods of 12-16 months, so only one application year. However, for more realistic probabilistic assessments a simulation period of about 20 years seems more appropriate. This will result in significantly increased simulation times. For the TOXSWA model simulation times may raise up to approximately 5 minutes for ponds, 15 minutes for ditches and one hour for streams. We investigated whether it would be possible to reduce the simulation time without compromising the accuracy of the predicted concentrations. In the current FOCUS scenarios TOXSWA uses segments of 30 m (ponds, 1 segment), 10 m (ditches, 10 segments) and 5 m (streams, 20 segments) in the numerical solution of the pesticide mass balance describing the concentration in the water layer. This allows to e.g. create concentration profiles as a function of distance in the ditches and streams. To reduce the simulation time rigorously, we cut down the number of segments in the water layer of ditches and streams to one segment. Next, concentrations calculated with a single segment for the water layer were compared to the maximum concentrations in the most downstream segment of the ditch or stream as used in the FOCUS scenarios. We considered maximum and 7 d time-weighted average concentrations both in water and sediment for a range of fictitious compounds. Initial simulations for the 12 and 16 months demonstrated that simulation times greatly reduced by replacing the standard FOCUS
segmentation in the water by a single water segment (still coupled to the standard 14 sediment segments). For the water layer we found that instantaneous peaks lowered up to 11% for ditches, but less than 2% for streams. For the sediment peak concentrations changed up to 20 or 30%, indiscriminately for ditches and streams. Consequently, time-weighted average concentrations also changed, up to 7% both in water and in sediment. Based upon these initial calculations replacing the standard 20 segments in FOCUS streams by one segment could be applied to obtain accurate concentrations in water, while significantly reducing simulation times.
MO123
Recent development of approaches for quantitative use of surface water monitoring data in aquatic exposure assessments
W. Chen, Syngenta Crop Protection, LLC; P. Mosquin, J. Aldworth, RTI International
Current pesticide regulatory ecological exposure assessments conducted by the U.S. Environmental Protection Agency are almost exclusively based on standard scenario computer modeling. Monitoring data may exist from targeted (prospective or retrospective) programs and/or general water quality research by industry, governments, and academic organizations. However, use of the monitoring measured data has been limited in the regulatory assessment process to
refine/inform modeling. The limited use of water monitoring data is largely due to variability in the monitoring program sampling designs (frequencies, timing etc.) and insufficient information regarding the exposure conditions and the context setting of the vulnerability of the monitoring location relative to a broader regional extent. In this paper, we summarize a set of recently developed approaches to infer and quantify realistic pesticide exposure potential based on monitoring data, including bias factor (BF), universal kriging (UK), and survey statistics. These approaches can be used in a systematic way to provide a useful reality check for comparison with exposure model output in regulatory assessments, thus increasing confidence in decision making. Examples of applying these approaches are provided to demonstrate their usefulness for watershed scale assessments.
Multi-year evaluations in the FOCUS Surface Water assessment - results of beta testing
D. Weber, M. Brauer, Eurofins Regulatory AG / Environmental and Ecological Modelling; A. Bolekhan, Bayer AG, Research & Development, Crop Science; G. Spickermann, ADAMA Deutschland GmbH; D. Schaefer, Bayer Crop Science / Environmental Safety
The calculations of the predicted environmental concentrations (PEC) of active substances in surface water are based on a “single year” approach with an initial 6 year warm-up phase followed by 16 months of the year selected by the FOCUS group. Unlike in groundwater with a 20 years assessment period, surface water exposure calculations based on a “single year” approach can be strongly affected by individual rainfall events (EFSA, 2013) which was discussed repeatedly by authorities, industry and academia (Klein, 2013, Goerlitz, 2015, Bach et al., 2016, Poulsen, 2016). This presentation provides background on the technical methods and assumptions currently implemented into a software tool (Weber et al., 2017) that allows 20-year simulations of FOCUS surface water scenarios. In addition, results of a beta test including revealed technical issues, problems and assumptions are discussed. The software tool in its current state can easily be adapted to updated technical requirements or changes, i.e. any comments from official side (EFSA FOCUS Repair Group) or from other sources can be addressed according to given consensus. The aim is to contribute to the development of an improved and generally accepted approach for surface water calculations representing a realistic worst case based on a robust evaluation. Bach M et al. (2016): Pesticide exposure assessment for surface waters in the EU. Part 1: Some comments on the current procedure. Pest Manag Sci 2016; 72: 1279–1284. April 2016 Goerlitz G (2015): Multiyear FOCUS surface water modelling: Options and Proposals for Realisation. XV: Symposium in Pesticide Chemistry. Piacenza, Italy. September 2015 Klein M (2013): Long term surface water simulations using the FOCUS scenarios. Pesticide Behaviour in oils, Water and Air, York, UK. September 2013 Poulsen V (2016): Higher tier assessments of aquatic and terrestrial studies. AGCHEM Forum, Barcelona. September 2016 Weber et al. 2017: Multi-Year evaluations in the FOCUS Surface Water assessment. Conference Pesticide Behaviour in Soils, Water and Air, York 2017.
MO125
Spatial and temporal explicit catchment modelling in aquatic risk assessment using the modular framework CMF
S. Multsch, F. Krebs, S. Reichenberger, DR. KNOELL CONSULT GmbH; S. Heine, Bayer Ag / Effect modelling; P. Kraft, L. Breuer, Justus Liebig University Giessen / Chair of Landscape, Water and Biogeochemical Cycles; T. Schad, Bayer Ag / Environmental Modelling
The EFSA Guidance Document on Aquatic Risk Assessment indicates a key role for effect modelling in future aquatic risk characterisation in a tiered risk assessment framework. Such approaches require correspondingly adapted exposure tools and scenarios ranging from simple edge-of-field to spatiotemporally explicit landscape-scale catchment models. These approaches should be sufficiently flexible and transparent in order to design lower- and higher-tiers of consistent protection level. Current models like SWAT or MIKE-SHE come with a fixed model structure which makes adaption to tiers of different complexity difficult. Flexible and modular approaches are needed to provide a spatially and temporally explicit aquatic exposure pattern to investigate effects on organisms according to Specific Protection Goals. A flexible and modular catchment model for water and pesticide transport has been developed which allows for stepwise adaption of model complexity to address tiered risk assessment problem formulations. The approach is based on the hydrological programming library CMF. Core functions of CMF are implemented in C++ and specific catchment setups are designed by Python scripting. The current approach focuses on the following abilities in order to investigate landscape-scale interactions: (a) a modular programming structure that enables replacement of process descriptions and (b) an incorporation of additional models; (c) an up-to-date connection between models at memory level in order to ensure high computing performance. A landscape is represented by the following components: Vertical water fluxes in fields are modelled with Richards equation, with the soil profile discretised into 24 soil layers. Each field holds a surface water, a groundwater and an optional drainage storage which are connected by a kinematic wave with the river bodies. Plant growth (phenology, leaf and root growth, water uptake) is modelled with similar methods and parameters as in the model MACRO 5.2. The setup was tested for a 350 ha catchment in Belgium under intensive arable use with detailed information on farming practice and observed discharge as well as herbicide loads at the catchment outlet for a time period of almost four years. The predicted environmental concentrations were used as input for an effect model in order to investigate the impact of the herbicide loads on the aquatic plant Lemna at population level.
MO126
Determination of runoff and drainage triggers for PEC surface water using automated simulation with FOCUS models
B. Kind, A. Guckland, J. Kleinmann, WSC Scientific GmbH
For the zonal registration in the EU predicted environmental concentrations in surface water need to be simulated based on the FOCUS models. Three different entry paths are considered: runoff (simulated in PRZM), drainage (MACRO) and
spray drift (SWASH drift calculator). While the latter only depends on the amount sprayed, the distance to the water body and the spray equipment used, runoff and drainage amounts are also triggered by substance properties, e.g. degradation in soil and adsorption to soil. Often, a lot of runs need to be simulated for different crops or application timings to proof a safe use of plant protection products as defined in the Good Agricultural Practice (GAP). For this poster we evaluate the FOCUS scenarios compared to the substance properties DT50 and KOC. The idea is to find DT50 and KOC values which trigger runoff and drainage events and to distinguish worst-case FOCUS scenarios for different DT50 and KOC values. Dummy substances will be created which have different values for KOC and/or DT50 in soil. The remaining properties will be identical for each KOC/DT50 variation. Using automated FOCUS surface water simulations PECsw values were calculated for different crops at different application times without consideration of spray drift as entry paths to focus solemnly on drainage and runoff. The results for different KOC/DT50 values of a single scenario were compared to identify a trigger value for runoff or drainage in this scenario. Furthermore, the results of different scenarios for a single substance will be compared to find the most sensitive scenario for these KOC/DT50 values. Finally, the amount of simulations necessary to show a safe use might be reduced to certain worst-case scenarios depending on the DT50 and KOC properties of the substance.
MO127
Quantitative exploitation of passive sampler data for pesticide mass flow calculation in catchments and exposure risk evaluation
T. Galle, Luxembourg Institute of Science and Technology; M. Bayerle, D. Pittois, V. Huck, Luxembourg Institute of Science and Technology LIST
Pesticide monitoring remains the blind spot in WFD monitoring schemes because of the episodic occurrence of their emissions following application periods. Full coverage of relevant exposure periods is logistically impossible on a larger scale with classical monitoring methods like grab or automatic sampling. Passive samplers provide a cost-effective solution that is deployable in great numbers allowing thereby a good spatial resolution. However, passive sampling still suffers from a lack of confidence of regulators and investigators with regard to the reliability of the ambient concentrations it represents and the supposed variability of sampling rates in the field. This contribution will show a rational monitoring strategy that has been applied in several catchments in Luxembourg and validated with parallel autosampling of flood events during application periods. It establishes that passive sampling is essentially time proportional and that base- and high flows can be separated for their contribution in terms of time-weighted averages and event mean concentrations. The biases and uncertainties in terms of load calculations are addressed. Based on monitoring in different hydrogeological contexts the approach is then used to derive land and crop use specific loads in catchments and exceedance probabilities of EQS values resulting in a risk map of impacted surface waters in Luxembourg.
MO128
Spatially distributed environmental fate modelling of terbuthylazine in a mesoscale agricultural catchment using passive sampler data
M. Gassmann, University of Kassel / Department Water Quality Management - Modelling and Simulation; T. Galle, Luxembourg Institute of Science and Technology; J. Farlin, Luxembourg Institute of Science and Technology LIST The impact of agricultural practices on water pollution can be assessed by process-based reactive transport modelling using catchment scale models. Most previous studies only used concentrations at the catchment outlet for model calibration and validation. Thus, even if the applied model is spatially distributed, predicted spatial differences of pesticide loss cannot be directly compared to observations. In this study, we applied the spatially distributed reactive transport model Zin-AgriTra in the mesoscale (78 km²) catchment of the Wark River in Luxembourg in order to simulate concentrations of terbuthylazine in river water. In contrast to former studies, we used six sampling points, equipped with passive samplers, for pesticide model validation. At each sampling point, event mean concentration of six events from May to July 2011 were calculated by subtraction of baseflow-mass from total collected mass assuming time-proportional uptake by passive samplers. Continuous discharge measurements and high-resolution autosampling during events allowed for accurate load calculations at the outlet. Detailed information about maize cultivation in the catchment and nation-wide terbuthylazine application statistics (average of 341 g/ha in the 3rd week of May) were used for a definition of the pesticide input function of the model. The hydrological model was manually calibrated to fit baseflow and spring/summer