Search

Machine Learning (ML) is increasingly applied to fill data gaps in assessments to quantify impacts associated with chemical emissions and chemicals in products. However, the systematic application of ML-based approaches to fill chemical data gaps is still limited, and their potential for addressing a wide range of chemicals is unknown. We prioritized chemical-related parameters for chemical toxicity characterization to inform ML model development based on two criteria: (1) each parameter's relevance to robustly characterize chemical toxicity described by the uncertainty in characterization results attributable to each parameter and (2) the potential for ML-based approaches to predict parameter values for a wide range of chemicals described by the availability of chemicals with measured parameter data. We prioritized 13 out of 38 parameters for developing ML-based approaches, while flagging another nine with critical data gaps. For all prioritized parameters, we performed a chemical space analysis to assess further the potential for ML-based approaches to predict data for diverse chemicals considering the structural diversity of available measured data, showing that ML-based approaches can potentially predict 8-46% of marketed chemicals based on 1-10% with available measured data. Our results can systematically inform future ML model development efforts to address data gaps in chemical toxicity characterization.

Pharmaceutical active compounds (PhACs) are a category of micropollutants frequently detected across integrated urban wastewater systems. Existing modelling tools supporting the evaluation of micropollutant fate in such complex systems, such as the IUWS_MP model library (which acronym IUWS stands for Integrated Urban Wastewater System), do not consider fate processes and fractions that are typical for PhACs. This limitation was overcome by extending the existing IUWS_MP model library with new fractions (conjugated metabolites, sequestrated fraction) and processes (consumption-excretion, deconjugation). The performance of the extended library was evaluated for five PhACs (carbamazepine, ibuprofen, diclofenac, paracetamol, furosemide) in two different integrated urban wastewater systems where measurements were available. Despite data uncertainty and the simplicity of the modelling approach, chosen to minimize data requirements, model prediction uncertainty overlapped with the measurements ranges across both systems, stressing the robustness of the proposed modelling approach. Possible applications of the extended IUWS_MP model library are presented, illustrating how this tool can support urban water managers in reducing environmental impacts from PhACs discharges.

The field of engineered nanomaterial (ENM) risk analysis has matured significantly in the past decade. While there is a suite of new, emerging tools to evaluate ENM risks and make decisions regarding these risks, there has not yet been thorough testing of these tools. This analysis applies and tests three risk screening tools (NanoRiskCat, LICARA nanoSCAN, NanoGRID) using a common case study focused on ENMs designed for water treatment technologies, compares results generated, and highlights key lessons learned and best practices for stakeholders involved in developing and/or applying ENM risk screening tools. NanoRiskCat was found to be most useful for providing a visual aid to characterize the potential exposure and health impact profiles of the ENMs, while LICARA nanoSCAN was most useful for providing guidance on proceeding with ENM-enabled innovations. NanoGRID was helpful for characterizing data on potential ENM exposure and hazards and providing detailed guidance for subsequent laboratory-based testing. At the same time, several key challenges were identified during tool application and testing phases, ranging from minor inconveniences to more complex, foundational issues. Key lessons learned and potential best practices gleaned from this analysis include: i) risk screening tools can be used together in a complementary manner; ii) risk managers and other users should be clear on the selection of underlying data and impacts on results; iii) multidisciplinary teams are essential for tool completion; and iv) continued testing and validation of emerging risk analysis tools for ENMs is a continued research need.

A new approach to automate the challenging task of defining the complexity level for conceptual modelling of micropollutant (MP) fate in sewer system is here presented. The approach combines GIS information and advanced statistical techniques (e.g., cluster analysis) and provides for a realistic description of in-sewer hydraulic residence time (HRT), which is fundamental for simulating MP removal processes occurring during in-sewer transport. The conceptual model was first tested in a full-scale catchment, where HRT distribution was determined based on spatial distribution of discharge sources and following calibration against high-frequency flow rate data. The model was then used to predict the dynamics of an ideal MP (biodegradation half-life = 2.5 h) at the outlet of the sewer system, revealing higher removal during in-sewer transport when considering an average HRT compared to a discrete HRT distribution. These results demonstrate that an intermediate complexity level, between highly detailed hydrodynamic models and simplified models, could be adopted for MP fate predictions while keeping computational demands reasonable. This latter aspect can be also of particular interest when an integrated modelling perspective (e.g., sewer and WWTP) is considered