Heterogeneity in biological assemblages and exposure in chemical risk assessment: Exploring capabilities and...

Chemical exposure concentrations and the composition of ecological receptors (e.g., species) vary in space and time, resulting in landscape-scale (e.g. catchment) heterogeneity. Current regulatory, prospective chemical risk assessment frameworks do not directly address this heterogeneity because they assume that reasonably worst-case chemical exposure concentrations co-occur (spatially and temporally) with biological species that are the most sensitive to the chemical's toxicity. Whilst current approaches may parameterise fate models with site-specific data and aim to be protective, a more precise understanding of when and where chemical exposure and species sensitivity co-occur enables risk assessments to be better tailored and applied mitigation more efficient. We use two aquatic case studies covering different spatial and temporal resolution to explore how geo-referenced data and spatial tools might be used to account for landscape heterogeneity of chemical exposure and ecological assemblages in prospective risk assessment. Each case study followed a stepwise approach: i) estimate and establish spatial chemical exposure distributions using local environmental information and environmental fate models; ii) derive toxicity thresholds for different taxonomic groups and determine geo-referenced distributions of exposure-toxicity ratios (i.e., potential risk); iii) overlay risk data with the ecological status of biomonitoring sites to determine if relationships exist. We focus on demonstrating whether the integration of relevant data and potential approaches is feasible rather than making comprehensive and refined risk assessments of specific chemicals. The case studies indicate that geo-referenced predicted environmental concentration estimations can be achieved with available data, models and tools but establishing the distribution of species assemblages is reliant on the availability of a few sources of biomonitoring data and tools. Linking large sets of geo-referenced exposure and biomonitoring data is feasible but assessment of risk will often be limited by the availability of ecotoxicity data. The studies highlight the important influence that choices for aggregating data and for the selection of statistical metrics have on assessing and interpreting risk at different spatial scales and patterns of distribution within the landscape. Finally, we discuss approaches and development needs that could help to address environmental heterogeneity in chemical risk assessment.