Worst-case ranking of organic substances detected in groundwater and surface waters in England

Spurgeon, David ORCID: https://orcid.org/0000-0003-3264-8760. 2021 Worst-case ranking of organic substances detected in groundwater and surface waters in England. Wallingford, UK Centre for Ecology & Hydrology, 96pp. (UKCEH Project no. 07514) (Unpublished)

Before downloading, please read NORA policies.

Preview

Text
N531239CR.pdf
Download (7MB) | Preview

Abstract/Summary

Background: The Environment Agency has been using GC-MS and LC-MS scans to semi-quantitatively measure organic substances in groundwater and surface waters. Lapworth et al. (2018) analysed this groundwater data to consider concentration ranges and spatial distribution. In this study, we extend this analysis to generate a worst-case hazard ranking of the detected substances in these groundwater samples and also in surface waters. It is intended that this ranking may be used to help identify substances for further consideration, e.g. for hazardous substance determinations under the Joint Agencies Groundwater Directive Advisory Group (JAGDAG), and/or through the Environment Agency’s Chemical Prioritisation and Early Warning System (PEWS). The results of ranking should not be used directly for control or management measures, due to the uncertainties inherent in the ranking process, for example, as a result of the semi-quantitative nature of the measurement data and preliminary nature of some of the hazard values used and also as it is a worst case ranking using the maximum concentration detected. The rankings are instead a means of identifying substance for further consideration. Overall approach: Accessible and downloadable hazard resources were used to collate hazard values for human health and ecological endpoints. UK Drinking Water Standards, and EFSA ADIs/TDIs were used in relation to human health and Water Framework Directive EQSs, NORMAN Network PNECs and chronic species sensitivity distribution (SSD) HC50 from Posthuma et al, 2019 for ecological hazard. The hazard values within each metric were compared to the highest measured concentration for each chemical to determine a hazard quotient. These hazard quotients were ranked for each of the human health metrics. For the three ecological hazard values, an average rank hazard rank was determined from the three metrics. These ranks were then multiplied by the substance detection frequency ranking to calculate an overall score for each chemical which was used as the final ranking of the substances in each media, i.e. surface water and groundwater, and for each analytical method, i.e. GC-MS and LC-MS. Use of a worst case approach, i.e. comparison with the highest detected concentration, was pragmatic as mean, median or 90th percentile values could not be estimated for most chemicals due to insufficient detections. To assess if the highest concentrations was an outlier, an assessment of this value in relation to other measured concentrations was conducted for 40 chemicals. Mixture effects were also considered using a concentration addition approach that assumes additivity of substance toxicity. Results: Substances detected using GC-MS and LC-MS screens were ranked for two human health metrics and for a combination of 3 ecological hazard values. Pesticides present in the top 30 ranked chemicals included legacy pesticides (particularly in groundwater) and current use actives (particularly in surface water). Intentional monitoring of specific uses were responsible for certain substances appearing in the top 30 ranked, for example rotenone which is used for invasive species control. A number of industrial and plastics associated chemicals were ranked highly in groundwater, while more consumer goods, personal care products and pharmaceuticals were ranked highly in surface waters. In all of the 40 individual substances cases assessed, the highest measured concentration was not found to be a substantial outlier. Both analysis methods identified the presence of complex mixtures in groundwater and surface water, although of lower complexity for GC-MS due to the higher detection limit. The GC-MS and LC-MS data and hazard metrics were used to generate a number of ranked lists of substances for future investigation. In developing the ranking approach, a number of decisions were made that could affect outcome, including hazard metric choice, metric weighting; hazard ranking correction by detection frequency, choice of detected concentration, assigning use categories and choice of mixture model. This final ranked list is not intended as a formal ranking of risk, but rather a prompt towards consideration for more detailed substance assessment within schemes such as JAGDAG and PEWS, as well as for other regulatory assessments and for designing research programs. The mixture assessment identified that cocktail effects can exceed those for any single chemical. Often, however, the magnitude of difference between predicted mixture risk effect and that for the most important single chemical was small. Indeed in >99% of all cases, the most toxic chemical contributed ≥ 20% of the mixture effect. This result demonstrates the feasibility of mixture assessment and the results are consistent with previous work (Backhaus and Faust, 2012).

Item Type:	Publication - Report (Technical Report)
UKCEH and CEH Sections/Science Areas:	Pollution (Science Area 2017-)
Funders/Sponsors:	Environment Agency
NORA Subject Terms:	Ecology and Environment
Date made live:	15 Oct 2021 14:11 +0 (UTC)
URI:	https://nora.nerc.ac.uk/id/eprint/531239

Item Type:

Publication - Report (Technical Report)

UKCEH and CEH Sections/Science Areas:

Pollution (Science Area 2017-)

Funders/Sponsors:

Environment Agency

NORA Subject Terms:

Ecology and Environment

Date made live:

15 Oct 2021 14:11 +0 (UTC)

URI:

https://nora.nerc.ac.uk/id/eprint/531239