Executive Summary

Models are used in exposure assessment for a number of reasons. They can help map the temporal and spatial variability of exposure, exposure pathways and exposure routes, and support risk assessment for water bodies where monitoring is lacking. They can be used to identify sources and pathways responsible for current exposures and to assess the impact of potential future developments of persistent, mobile, and toxic chemicals (PMT) exposures in surface water and groundwater. Such scenario assessment may include changes in PMT use, effects of pollution control measures, accidental spills or climate change.

The scope of this document, produced as part of the H2020 PROMISCES project, is to provide guidance for applications of models with a specific focus on model trains for the assessment of exposure to PMTs as part of the predictive risk assessment related to surface and groundwater. This document explains the basic concepts of specific models and how best to use them in model
trains in the framework of a tiered approach. The intention is to inform users and interested stakeholders about what needs to be considered when using different methods, what is the best use of specific models, what are the best combinations in model trains and what are their current limitations.

The guidance document presents (i) “screening level” models for the assessment of regional exposure of groundwater from soil pollution and for the assessment of general exposure of air, soil and water at local, regional or global scales, (ii) spatial and temporal explicit approaches for the identification of pollution plumes in the soil-groundwater continuum and (iii) model train applications for the catchment – river – river bank filtration – drinking water continuum.

Exposure of surface water and groundwater to PMT depends on the use patterns and the environmental fate of the chemicals. Emission, fate and transport models incorporate driving factors into documented algorithms. The extent to which a substance persists in surface water can, for instance, be calculated with the “SimpleBox - Aquatic Persistence Dashboard”, based on its physical-chemical characteristics. The presented approach for deriving generic risk limits for soils shows that, depending on regional variations in geo(hydro)logical conditions, the high mobility of some PFAS could lead to strict requirements for materials applied on soil.

For the soil-groundwater continuum, a novel model train is presented which accounts for the main physical and chemical processes controlling the fate and transport of PFAS. For sorption and degradation reactions, several formalisms can be used, allowing one to select the most appropriate according to the PFAS molecular properties and the characteristics of the simulated
domain. The results issued from these modelling applications indicate the key role of correctly identifying the main physical, chemical and biological processes controlling fate and transport of PFAS in the studied domain to build a robust conceptual model. To increase the robustness of the model, a thorough model calibration must be performed, preferably using time series
measurements of the PFAS concentration in the pore solution at different locations of the contaminated site.

The results confirm the key role of the unsaturated zone in the transfer and long-term migration of PFAS. Nonlinearity and nonideality of sorption reactions were expected for a broad range of PFAS, suggesting using more complex numerical formalism than linear isotherms. Considering the key role of capillary fringe displacement on PFAS transport in the unsaturated zone, the
model train seems to be very efficient in performing PFAS simulations, as it can explicitly describe water flow and solute transport at the interface between the unsaturated and saturated zones, avoiding the main pitfall encountered in other numerical approaches.

The combination of stand-alone models in model trains expands the scope that can be covered in the context of a catchment – river – riverbank filtration – drinking water continuum for exposure assessment of surface waters and bank filtered drinking water. Model trains can combine individual models either in a complementary way or in a sequence. A complementary combination may either compare models of different complexity to find out which level of complexity (and associated effort) is needed to answer which questions, or may compare different models with their different strengths and weaknesses in parallel to assess uncertainties and/or use models for scenario evaluation according to their specific capabilities. A sequential combination facilitates a broader application in terms of content and at different spatial resolutions. Clearly defined interfaces are essential for a successful implementation.

Examples of model trains for selected PFAS are presented for the catchment-river interaction in the urban context of the Berlin case and for the whole catchment – river – riverbank filtration – drinking water continuum on the scale of the Upper Danube Basin. The Berlin case demonstrates the application of the sequential model train by combining a city emission model with a city surface water fate and transport model to assess the resulting exposure to PFAS in the city surface waters. The Danube case demonstrates the application of a sequential model train for exposure assessment of bank filtered drinking water by combining large-scale catchment-scale emission models with different types of bank filtration fate and transport models for specific locations in the catchment. In addition, it also demonstrates complementary application by comparing emission models with different strengths and weaknesses for the assessment of multiple scenarios on the catchment scale and different levels of complexity for the fate and transport modelling of bank filtration. The model train has been successfully applied for 10 different PFAS-substances including the assessment of a large range of scenarios.

Current limitations for exposure assessment of PFAS at river basin scale require improvement in scientific understanding as well as additional efforts in administrative data collection and inventory development. Current results of the exposure assessment show the very high relevance of legacy pollution from use of fire-fighting foams or from old municipal landfills. On the administrative level, there is a strong need for improved identification and harmonized inventorying of contaminated sites at national and international (EU) level. The lack of robust, openly available information on production, import-export and therefore use volumes of PFAS at national and EU level is strongly hampering exposure assessment. A major effort is urgently needed to provide this information, as it is decisive for a sound environmental exposure assessment, not only for surface water and groundwater.

In regard to scientific advances, there is a need for more and better understanding of the extent of local groundwater pollution, particularly due to the application of fire-fighting foams or to the presence of municipal landfills. Further improvement of the scientific knowledge about the fate of PFAS in the environment, including their partitioning between different phases (air,
water, solids) and the transformation of the so called “precursors” into stable “end-products” like PFOA, PFOS and short-chain substances is needed to enlarge the number of PFAS that can be included into the exposure assessment. A reproducible and standardised analytical parameter for “total PFAS” or even “total toxicity of PFAS” would be needed to address all relevant PFAS in a combined way as it is a focus of Workpackage 1 of the H2020 PROMISCES project (Togola et al. 2024; Behnisch et al. 2024).