The effect of intervention in nickel concentrations on benthic macroinvertebrates: A case study of statistical causal inference in ecotoxicology

A modelling framework to assess multiple metals impacts on marine food webs: Relevance for assessing the ecological implications of deep-sea mining based on a systematic review

Industrial deep-sea mining will release plumes containing metals that may disperse over long distances; however, there is no general understanding of metal effects on marine ecosystems. Thus, we conducted a systematic review in search of models of metal effects on aquatic biota with the future perspective to support Environmental Risk Assessment (ERA) of deep-sea mining. According to results, the use of models to study metal effects is strongly biased towards freshwater species (83% freshwater versus 14% marine); Cu, Hg, Al, Ni, Pb, Cd and Zn are the best-studied metals, and most studies target few species rather than entire food webs. We argue that these limitations restrain ERA on marine ecosystems. To overcome this gap of knowledge, we suggest future research directions and propose a modelling framework to predict the effects of metals on marine food webs, which in our view is relevant for ERA of deep-sea mining.

Evaluating the Protectiveness of a Bioavailability-Based Environmental Quality Standard for the Protection of Aquatic Communities from Zinc Toxicity Based on Field Evidence

Environmental quality standards (EQS) are typically derived from the results of laboratory studies on single species. There is always uncertainty surrounding the protectiveness of an EQS when applied to real ecosystems containing a multitude of chemical and physical stressors. Quantile regression was used with field biological data on invertebrates in United Kingdom waters to identify taxa that are responsive to bioavailable zinc exposures. A threshold based on the total abundance of eight responsive taxa is used as an indicator of the overall ecosystem sensitivity. The inclusion of some responsive but insensitive taxa in this ecological metric could bias the results toward a higher threshold. The least responsive species were progressively removed from the collective ecological metric, basing the analysis on a progressively smaller number of the more responsive species. Quantile regression analysis at the 95th quantile for the three most responsive taxa resulted in a 10% effect concentration of 14.8 µg L^-1 bioavailable zinc, suggesting that the EQS of 10.9 µg L^-1 bioavailable zinc is sufficiently protective of sensitive members of the invertebrate community. There is a compromise between the robustness of the analysis and the sensitivity of the subcommunity that it is based on. Analyses based on fewer taxa provide a more sensitive result. This approach assessed real ecosystem data and evaluated the uncertainty associated with the protectiveness of the EQS for zinc. The zinc EQS is sufficiently protective of sensitive members of benthic macroinvertebrate communities under real environmental conditions, including a mix of multiple substances. Environ Toxicol Chem 2023;42:1010-1021. © 2023 The Authors. Environmental Toxicology and Chemistry published by Wiley Periodicals LLC on behalf of SETAC.

What do we want to estimate from observational datasets? Choosing appropriate statistical analysis methods based on the chemical management phase

The goals of observational dataset analysis vary with the management phase of environments threatened by anthropogenic chemicals. For example, identifying severely compromised sites is necessary to determine candidate sites in which to implement measures during early management phases. Among the most effective approaches is developing regression models with high predictive power for dependent variable values using the Akaike information criterion. However, this analytical approach may be theoretically inappropriate to obtain the necessary information in various chemical management phases, such as the intervention effect size of a chemical required in the late chemical management phase to evaluate the necessity of an effluent standard and its specific value. However, choosing appropriate statistical methods based on the data analysis objective in each chemical management phase has rarely been performed. This study provides an overview of the primary data analysis objectives in the early and late chemical management phases. For each objective, several suitable statistical analysis methods for observational datasets are detailed. In addition, the study presents examples of linear regression analysis procedures using an available dataset derived from field surveys conducted in Japanese rivers. Integr Environ Assess Manag 2022;18:1414-1422. © 2021 The Authors. Integrated Environmental Assessment and Management published by Wiley Periodicals LLC on behalf of Society of Environmental Toxicology & Chemistry (SETAC).

What drives the ecological quality of surface waters? A review of 11 predictive modeling tools

What policy is needed to ensure that good-quality water is available for both people's needs and the environment? The EU Water Framework Directive (WFD), which came into force in 2000, established a framework for the assessment, management, protection and improvement of the status of water bodies across the European Union. However, recent reviews show that the ecological status of the majority of surface waters in the EU does not meet the requirement of good status. Thus, it is an important question what measures water management authorities should take to improve the ecological status of their water bodies. To find concrete answers, several institutes in the Netherlands cooperated to develop a software tool, the WFD Explorer, to assist water managers in selecting efficient measures. This article deals with the development of prediction tools that allow one to calculate the effect of restoration and mitigation measures on the biological quality, expressed in terms of Ecological Quality Ratios (EQRs). To find the ideal modeling tool we give a review of 11 predictive models: 10 models from the field of Machine Learning and, additionally, the Multiple Regression model. We present our results in terms of a 'prediction-interpretation competition'. All these models were tested in a multiple-stressor setting: the values of 15 stressors (or steering factors) are available to predict the EQR values of four biological quality elements (phytoplankton, other aquatic flora, benthic invertebrates and fish). Analyses are based on 29 data sets from various water clusters (streams, ditches, lakes, channels). All 11 models were ranked by their predictive performance and their level of model transparency. Our review shows a trade-off between these two aspects. Models that have the best EQR prediction performance show non-transparent model structures. These are Random Forest and Boosting. However, models with low prediction accuracies show transparent response relationships between EQRs on the one hand and individual steering factors on the other hand. These models are Multiple Regression, Regression Trees and Product Unit Neural Networks. To acknowledge both aspects of model quality - predictive power and transparency - we recommend that models from both groups are implemented in the WFD Explorer software.

Application of Bioavailability Models to Derive Chronic Guideline Values for Nickel in Freshwaters of Australia and New Zealand

There has been an increased emphasis on incorporating bioavailability-based approaches into freshwater guideline value derivations for metals in the Australian and New Zealand water quality guidelines. Four bioavailability models were compared: the existing European biotic ligand model (European Union BLM) and a softwater BLM, together with 2 newly developed multiple linear regressions (MLRs)-a trophic level-specific MLR and a pooled MLR. Each of the 4 models was used to normalize a nickel ecotoxicity dataset (combined tropical and temperate data) to an index condition of pH 7.5, 6 mg Ca/L, 4 mg Mg/L, (i.e., approximately 30 mg CaCO₃ /L hardness), and 0.5 mg DOC/L. The trophic level-specific MLR outperformed the other 3 models, with 79% of the predicted 10% effect concentration (EC10) values within a factor of 2 of the observed EC10 values. All 4 models gave similar normalized species sensitivity distributions and similar estimates of protective concentrations (PCs). Based on the index condition water chemistry proposed as the basis of the national guideline value, a protective concentration for 95% of species (PC95) of 3 µg Ni/L was derived. This guideline value can be adjusted up and down to account for site-specific water chemistries. Predictions of PC95 values for 20 different typical water chemistries for Australia and New Zealand varied by >40-fold, which confirmed that correction for nickel bioavailability is critical for the derivation of site-specific guideline values. Environ Toxicol Chem 2021;40:100-112. © 2020 The Authors. Environmental Toxicology and Chemistry published by Wiley Periodicals LLC on behalf of SETAC.