Outcomes from an ECETOC task force on geospatial approaches to increasing the ecological relevance of chemical risk assessments
SETAC Europe 2019, Poster
Session Title: 4D-Risk Assessment
Monday, May 27, 2019
5:15 pm, Exhibition Hall
For several decades, the prospective risk assessment of chemicals has followed a generic approach of comparing estimated exposures to toxic thresholds designed to be protective of all species (i.e., assessing exposure to the most sensitive species assumed to be located anywhere the chemical may occur in the environment). This approach does not recognise geographic patterns of species distributions or acknowledge that particularly sensitive species may not occupy potentially exposed habitats. Therefore, risk assessments could be overly conservative and restrictive for some uses of chemicals. Approaches for making spatially explicit assessments of chemical exposure are relatively advanced but this is not the case for mapping and assessing ecological data. However, geo-referenced ecological data appear to be increasingly available at spatial resolutions applicable to chemical risk assessment, potentially facilitating enhanced environmental relevance of such risk assessments. In 2017 a Task Force was initiated by European Centre for Ecotoxicology and Toxicology of Chemicals (ECETOC) to assess the capability of making chemical risk assessments using available geospatially referenced chemical exposure, ecological receptor and ecosystem services data. Two case studies were developed to illustrate the potential to assess geo-referenced risks to ecological receptors in fresh water and terrestrial environments exposed to i) a chemical used in consumer cleaning products discharged via municipal WWTPs across the EU and ii) a range of representative active ingredients used in plant protection products on selected crops in Germany. After initially compiling a catalogue of available geo-referenced ecological data for Europe, geo-referenced exposure concentrations were derived by combining accessible chemical use and fate data with conventional exposure models. However, use of many ecological data sets over a pan-European range proved problematic due to data access issues, limited geographic coverage and unreliable quality. Nevertheless, several suitable ecological data sets were accessed after making specific requests to various organisations within national authorities and these were integrated with the exposure maps. The results of these case studies give an indication of the potential value of making geo-referenced chemical risk assessments as well as the limitations to current capability.
S. Marshall (Consultant), A. Ireland (ExxonMobil Biomedical Sciences), J.C. Otte (BASF), L. Maltby (presenter; University of Sheffield), C. Holmes (Waterborne Environmental), and P. Sweeney (Syngenta). Outcomes from an ECETOC task force on geospatial approaches to increasing the ecological relevance of chemical risk assessments. Poster SETAC Europe 2019. Helsinki, Finland.
Applying the mechanistic honey bee colony model BEEHAVE to assess multiple factors impacting overwintering survival in large colony feeding studies
SETAC Europe 2019, Poster
Session Title: Bees, bugs and beneficials in environmental risk assessment and testing
Tuesday, May 28, 2019
8:30 am, Exhibition Hall
Honey bee Large Colony Feeding Study (LCFS) is a novel type of Tier II semi-field study for the determination of potential effects of pesticides on free-foraging whole colonies during and after dietary intake of a known pesticide concentration. LCFS are currently accepted by North American regulatory agencies and represent a progressively more realistic level of refinement compared to individual laboratory-based studies. LCFS are designed to test toxicity via consumption of fed sucrose syrup over a worst case exposure scenario of six weeks, with colony assessments conditions over a foraging season and following overwintering period. However, such studies are very cost- and time-intensive, and high overwintering losses of control hives have been observed in some studies. Loss of control colonies indicates that stressors other than pesticides, e.g. resource availability, weather, diseases and beekeeping activities, likely influence colony overwintering survival, confounding the assessment of impacts caused by pesticides. In the current study commissioned by the Pollinator Research Task Force, we apply the mechanistic honey bee colony model BEEHAVE to simulate colony dynamics observed in negative control colonies from multiple colony feeding studies. Detailed landscape-level data inform the resource availability for the simulated foragers in the model. In addition, weather data, initial colony condition and feeding patterns were analyzed across studies and translated to model inputs. In a calibration step, we adjusted parameters in BEEHAVE to achieve simulated dynamics corresponding to colony conditions reported in the studies. Study data collected in summer and fall were analyzed for predictors of overwintering success of individual colonies. BEEHAVE simulations with different combinations of external factors were used to assess their importance for colony condition. Colony conditions at study initialization and feeding patterns both influenced the colony condition in the fall, and thus, the probability of overwintering survival. Model simulations with different colony feeding patterns and initial colony conditions were then used to quantitatively estimate colony-level outcomes under conditions deviating from those in the studies. These results can be used to improve and inform LCFS study designs. Pesticide effects can be included in future model analyses, and analyzed in the context of multiple factors that impact colony health and overwintering success.
Amelie Schmolke (Waterborne Environmental), Farah Abi-Akar (Waterborne Environmental), Nika Galic (Syngenta), Silvia Hinarejos (presenter; Sumitomo Chemical Co.). Applying the mechanistic honey bee colony model BEEHAVE to assess multiple factors impacting overwintering survival in large colony feeding studies. Poster SETAC Europe 2019. Helsinki, Finland.
Modelling emissions of microplastics in Europe from wastewater sources, including land applied biosolids
Public information regarding microplastics in the environment is frequently available and comes from a variety of sources, often in the form of retrospective sources such as measured aquatic data. Science-based risk assessment must utilize both retrospective and prospective exposure information to effectively estimate potential risk to ecological receptors. While monitoring data provide information at only a few locations for several points in time, prospective models can estimate the potential for ecological exposures across many landscapes and over long periods of time, and both have a role in risk assessment. Wastewater treatment plants are often cited as a source of microplastics entering the environment. Microplastics are highly removed (generally >90%) during the waste water treatment process, via skimming of floating particles or sorption to solids and settling into sludge. Understanding the eventual fate of this sludge, and the potential for contained microplastics to re-enter surface water, is one step of many in determining the fate of microplastics in the aquatic environment. Sludge management in Europe varies geographically, with up to 90% of sludge used on agriculture in Portugal, and 0% in other countries (Eurostat, 2017) with other disposal including incineration, landfill or composting. We present a model which addresses both direct aquatic emissions into surface water via waste water effluent, as well as indirectly from land applied biosolids coupled with spatially-defined surface runoff potential. Generalized runoff potential is estimated using fate and transport models used for plant protection products found in the EFSA FOCUS scenarios. To our knowledge, this coupling of direct aquatic emission and sludge-biosolids-runoff is a novel approach for examining environmental emissions of microplastics which enter municipal wastewater treatment plants. This spatially-explicit model is based on publicly available datasets, combined with a hydrologic framework containing geographically variable emissions linked to a river network simulating environmental transport via surface water.
Christopher Holmes, Joshua Amos, Amy Ritter, Marty Williams, and Scott Dyer (Waterborne Environmental). Modelling emissions of microplastics in Europe from wastewater sources, including land applied biosolids. Poster SETAC Europe 2019. Helsinki, Finland.
A prospective approach for assessing chemical mixtures in river catchments with diverse land uses
Field-based ecological risk assessments incorporate risks from chemical mixtures and a myriad of stressors because ecosystems are continuously exposed to a wide-array of contaminants and nonchemical stressors. Considering the large numbers potential combinations of mixtures and stressors, this problem could seem insurmountable. We demonstrate that such combinations can be simplified by 3 land-use related chemical emission scenarios: agriculture, domestic, and urban. We applied a tiered methodology to assess the implications of each of the scenarios via a quantitative model. The results showed land use–dependent mixture exposures, clearly discriminating downstream effects of land uses, with unique chemical “signatures” regarding composition, concentration, and temporal patterns. Associated risks were characterized in relation to the land-use scenarios. Comparisons to measured environmental concentrations and predicted impacts showed relatively good similarity. The results suggest that the land uses imply exceedances of regulatory protective environmental quality standards, varying over time in relation to rain events and associated flow and dilution variation. Higher-tier analyses using ecotoxicological effect criteria confirmed that species assemblages may be affected by exposures exceeding no-effect levels and that mixture exposure could be associated with predicted species loss under certain situations. The model outcomes inform various types of prioritization to support risk management, including a ranking across land uses as a whole, a ranking on characteristics of exposure times and frequencies, and various rankings of the relative role of individual chemicals. Though all results are based on in silico assessments, our land use–based approach yields useful insights for simplifying and assessing potential ecological risks of chemical mixtures and can therefore be useful for catchment-management decisions.
Scott Dyer (Waterborne Environmental), Leo Posthuma (RIVM), Colin D. Brown (University of York), Dick de Zwart (RIVM), Jerome Diamond (Tetra Tech), Christopher Holmes (Waterborne Environmental), Stuart Marshall (Bedford, UK), and G. Allen Burton Jr (University of Michigan). A prospective approach for assessing chemical mixtures in river catchments with diverse land uses.
Poster SETAC Europe 2019. Helsinki, Finland.
Applying a mechanistic honey bee colony model to assess multiple factors impacting colony overwintering survival
Honey bee colony feeding studies are one type of Tier II semi-field studies designed to determine potential effects of pesticides on free-foraging whole colonies during and after dietary intake of a known pesticide concentration. These studies represent a progressively more realistic level of refinement for pollinator studies compared to individual laboratory-based studies since they are intended to reflect a worst-case exposure scenario in the field. Colony feeding studies are designed to test toxicity over a foraging season and following overwintering period. However, such studies are very cost- and time-intensive to conduct, and high overwintering losses of control hives have been observed in some studies. Loss of control colonies indicates that stressors other than pesticides, e.g. resource availability, weather, diseases and beekeeping activities, likely influence colony overwintering survival, confounding the assessment of impacts caused by pesticides. Honey bee colony models have been gaining interest as tools in pesticide risk assessment to inform study design and ultimately, colony-level risks to honey bees. In the current study commissioned by the Pollinator Research Task Force, we apply the honey bee colony model BEEHAVE to simulate colony dynamics observed in negative control colonies from multiple colony feeding studies. Detailed landscape-level data inform the resource availability for the simulated foragers in the model. In addition, weather data, initial colony condition and feeding patterns were analyzed across studies and translated to model inputs. In a calibration step, we adjusted parameters in BEEHAVE to achieve simulated dynamics corresponding to colony conditions reported in the studies. Study data collected in summer and fall were analyzed for predictors of overwintering success of individual colonies. BEEHAVE simulations with different combinations of external factors were used to assess their importance for colony condition. Colony conditions at study initialization and feeding patterns both influenced the colony condition in the fall, and thus, the probability of overwintering survival. Model simulations can be used to estimate colony-level outcomes under conditions deviating from those in the studies to inform study design and extend the use of the available data. Pesticide effects can be included in future model analyses, and analyzed in the context of multiple factors that impact colony health and overwintering success.
Amelie Schmolke (Waterborne Environmental), Farah Abi-Akar (Waterborne Environmental), Nika Galic (Syngenta), Silvia Hinarejos (Sumitomo). Applying a mechanistic honey bee colony model to assess multiple factors impacting colony overwintering survival. Platform SETAC 2018. Sacramento, CA.
Influence of particle size on prospectively modeled environmental concentrations of microplastics in the Sandusky River watershed
The presence of nano- and microplastics (MPs; particles < 5 mm) in the aquatic environment is a topic of increasing discussion and research. Although measurement and monitoring data are indispensable, there is a need to prospectively estimate concentrations to enable forward-looking assessments and to guide analysis of retrospective ecological analyses. For traditional chemicals, fate and exposure models have been proven to be very helpful and are widely used. However, to date few models exist that simulate the transport and fate of MPs in freshwater systems. This presentation presents simulations of the transport and fate of various-sized MPs emitted from wastewater treatment plants into freshwater riverine systems, and tracks concentrations moving downstream from headwater into Lake Erie. We linked the NanoDUFLOW model (a detailed MP aggregation-sedimentation model integrated in a hydrological and particle transport model) with iSTREEM® (developed to estimate chemical concentration distributions for all rivers receiving WWTP discharges in the US) for a range of particle sizes. This combines the mechanistic realism of NanoDUFLOW, accounting for formation and settling of heteroaggregates, with the US well-established iSTREEM implementation. Depth dependent in-stream first order removal rate constants simulated with NanoDUFLOW were combined with standard iSTREEM output which simulated the emission, transport and water column concentrations of different MP sizes. We modeled floating as well as non-buoyant MP, for sizes ranging from 100 nm to 1000 µm. We also modeled a combined mixture of particle sizes based on effluent measurements from Mason et al (2016). Simulations were spatially explicit with MP concentrations being modeled for the Sandusky River watershed in Ohio containing over 300 miles of river downstream of 20 WWTPs. Modelling results show the effects of population density, MP size and environmental conditions on riverine concentrations and export to Lake Erie. Buoyant as well as the smallest non-buoyant MP fractions can be transported over long distances, reaching receiving waters such as the Great Lakes. In contrast, larger non-buoyant MPs settle more locally in the vicinity of the WWTPs.
Christopher Holmes (Waterborne Environmental), Albert Koelmans (Wageningen University), Scott Dyer (Waterborne Environmental). Influence of particle size on prospectively modeled environmental concentrations of microplastics in the Sandusky River watershed. Poster SETAC 2018. Sacramento, CA.
Spatial and temporal variations of national cropping patterns for higher-tier pesticide exposure assessment
Pesticides are used on numerous agricultural crops across the United States to control pests and improve food yield and quality. This presentation focuses on the spatial and temporal aspects of a national scale assessment conducted by the Pyrethroid Working Group (PWG) to characterize the potential for pyrethroids to enter flowing surface waters based on a spatially explicit analysis of crop proximity to surface waters using multi-year data on diverse agricultural production patterns. Standard exposure calculations in the USEPA EFED regulatory risk assessment framework assume that 100% of the area around the water body (the Tier II pond) is cropped and treated, and therefore subject to drift and runoff entry. Over two million catchments within the National Hydrography Dataset (NHDPlus) were characterized using geospatial data to develop national or regional metrics related to potential surface water exposure related to crop proximity for more than 10 crop types. Crop locations were based on five years of data from USDA NASS Cropland Data Layer. Results highlight the variability of cropping density at the catchment scale across different geographies and scales, as well as situations in which density of potentially highly exposing crop (e.g., within 200m of surface water) may not match ‘entire catchment’ cropping density patterns. Variations in cropping density (as a proxy for potential exposure) across multiple years will be discussed in relation to how this variability may influence exposure estimates. When examining all catchments containing a specific crop, the 90th percentile crop density values (based on the 200m proximity zone) ranged from 1.3% (for vegetables in FL) to 44.4% (for tree nuts in CA). The resulting datasets provide a useful set of metrics across multiple crops which can be applied to pesticide exposure assessments that may need spatially-explicit refinements related to crop-water interactions. Because the crop proximity results are linked to the NHD+ framework, including these specific attributes into other NHD+-based analyses is extremely efficient.
Christopher Holmes, Joshua Amos (Waterborne Environmental), Paul Hendley (Phasera, Ltd.), Russell Jones (Bayer CropScience), Scott Jackson (Valent). Spatial and temporal variations of national cropping patterns for higher-tier pesticide exposure assessment. Poster SETAC 2018. Sacramento, CA.
A generalized life-history model for assessing indirect effects of pesticides on fish populations
Assessing population-level effects of stressors, such as pesticides, across species is challenging because effects are influenced not only by individual-level toxicity but also by species’ life history characteristics, ecology, and the duration, magnitude, and frequency of stressor exposure. Additionally, potential indirect population-level effects of stressor exposure (for example, effects resulting from changes in food availability) add further complexity to the assessment. To address these challenges, we developed a generalized population model for small fishes for assessing potential indirect effects of pesticide exposure on population dynamics. We applied the model to 17 species of darter (Percidae: Etheostomatinae) for which life history and diet data were readily available. The selected darter species (including two species listed under the Endangered Species Act) cover a range of life history strategies and diet compositions within the subfamily. We simulated several scenarios of pesticide exposure to assess how potential impacts of a pesticide on various prey species (invertebrates) included in the diets of selected darter species might affect darter population dynamics over extended time periods. We also investigated correlations between life history characteristics and population responses to the exposure scenarios. This analysis provides a framework for predicting food web mediated effects of pesticides on darter species for which little life history information is available. By combining life history variability in fish with estimates of potential indirect effects of pesticides on their prey, our model can provide a valuable tool for incorporating ecological complexity into the assessment process to quantify population-relevant risks to listed and non-target species of small fish.
Nicholas Green, Amelie Schmolke, Brian Kearns, Colleen Roy, Katherine Kapo, Matthew Kern (Waterborne Environmental), Alan Samel (FMC), Valery Forbes (University of Minnesota). A generalized life-history model for assessing indirect effects of pesticides on fish populations. Poster SETAC 2018. Sacramento, CA.
Development of a spatially resolved global mean annual flow dataset for use in environmental risk assessment: A case study for China
Environmental exposure models for chemicals used widely across large geographic areas and disposed of down the drain are important tools for informing ecological risk assessments. One important element of these models is understanding the dilution of wastewater treatment plant (WWTP) effluent into the receiving stream (dilution factors) which allows for the estimation of in-stream environmental concentrations based on either estimated flow of receiving waters. In the U.S., the iSTREEM model (American Cleaning Institute) estimates dilution of WWTP effluent into receiving streams through the incorporation of a spatial hydrologic network with associated flow data (National Hydrography Dataset Plus) into the exposure model to spatially associate (and route) local WWTP emissions with corresponding local flows. A similar approach for generating localized dilution factors can be employed on the global scale to integrate the chemical emissions component of the model with a hydrologically-connected global river network with associated flow values. The HydroSHEDS and HydroBASINS datasets (Lehner et al. 2008 and 2013) provide a global hydrology dataset that can be used as a spatial hydrologic framework, including a network of streams and rivers and watershed and catchment boundaries. However, flow estimates corresponding to the global river network are a critical attribute that must still be incorporated for exposure modeling. Using China as a case study, a mean annual flow dataset to correspond with the HydroSHEDS and HydroBASINS global data was developed using the well-established Curve Number (CN) approach developed by Natural Resources Conservation Service (NRCS, USDA). The CN approach integrates environmental and landscape features including best available and high-resolution precipitation, soils, and land use characteristics to estimate surface runoff over the land area. The high-resolution runoff grid was spatially combined with hydrology datasets to derive flow estimates across a river network. Global datasets were utilized for model parameters so that the approach could be extrapolated to the global scale, while also providing the flexibility to incorporate best-available data. This presentation will provide a detailed overview of the runoff methodology, validation against measured flow data, and the resulting river flow dataset for China.
Raghu Vamshi, Katherine Kapo, Amy Ritter, Brian Kearns (Waterborne Environmental), and Kathleen McDonough (Procter & Gamble). Development of a spatially resolved global mean annual flow dataset for use in environmental risk assessment: A case study for China. Poster SETAC 2018. Sacramento, CA.
Development of a global environmental exposure modeling framework for risk assessment of chemicals disposed down the drain: A case study for China
Environmental exposure assessment of chemicals that are disposed down the drain (such as consumer product ingredients) at the global scale within a consistent and accessible framework has remained a challenge over the years, despite advancements in exposure modeling and global and local data resources. Historically, assessment efforts have been tailored and applied to specific geographies and used simplistic approaches rather than to build a spatially resolved global assessment infrastructure. Challenges such as inconsistent, scarce, or rapidly-evolving data resources, particularly for developing countries where assessment needs are high, have further complicated the evolution of spatially resolved global exposure assessment tools. However, through strategic integration of existing global data resources and established modeling tools, a standardized framework and methodology for GIS-based exposure modeling can be developed for the global scale. In this study, we present a spatially resolved global environmental exposure model approach designed to incorporate best-available data and modeling tools, using China as a case study. The global hydrology network from HydroSHEDS and HydroBASINS (Lehner et al. 2008 and 2013), global river flow and population estimates, and best-available country-specific water use and wastewater treatment information were integrated with the GIS-ROUT exposure model (Wang et al. 2005) and iSTREEM® model framework (American Cleaning Institute) to provide a means of estimating the distribution of concentrations of a chemical disposed down the drain across a river network based on chemical production volume and consumer usage estimates. Both wastewater treatment plant effluent and direct discharge are accounted for by the model through estimation of catchment-specific emissions. The spatial nature of the model provides a robust means for estimating variability in environmental exposures. Details of the various model components and generated output for China are overviewed, as well as considerations and discussion regarding on-going extrapolation to the global scale. The framework developed as part of this model is highly adaptable to countries with an abundance of data (e.g., North America, Western Europe, etc.) or those scarce with data (e.g., developing countries) available to parametrize the model.
Kathleen McDonough (Procter & Gamble), Katherine Kapo (Waterborne Environmental), and Raghu Vamshi (Waterborne Environmental). Development of a global environmental exposure modeling framework for risk assessment of chemicals disposed down the drain: A case study for China. Poster SETAC 2018. Sacramento, CA.