Frederik Verdonck

Environmental risk assessment of zinc in European freshwaters: A critical appraisal

A risk assessment report (RAR) on zinc and zinc compounds has recently been prepared in the framework of the European Union (EU) Council Regulation 793/93/EEC on Existing Chemicals. The EU Scientific Committee on Human and Environmental Risks (SCHER) has, however, expressed some fundamental, science-based concerns about the approach followed and the conclusions. The main objective of the present study was to assess the potential environmental risks associated with current use patterns of Zn in nine EU river basins in Germany, France and Belgium, thereby using more advanced methodologies which are largely in line with the recommendations made by SCHER. This included (i) avoiding working with measured Zn concentrations from monitoring stations that were potentially influenced by point sources and/or historical contamination, (ii) the full bioavailability normalization of all chronic ecotoxicity data to river basin specific physico-chemistry using biotic ligand models (BLM), prior to deriving predicted no effect concentrations (PNEC) with the species sensitivity distribution (SSD) approach, and (iii) the use of a probabilistic framework for risk characterization. Further, a total risk approach instead of an added risk approach was used, and the PNEC was equated to the HC5-50 without an additional assessment factor. Based on monitoring data we estimated predicted environmental concentrations (PEC) for the different EU river basins between 1.3 and 14.6 microg dissolved Zn/L. PNEC values varied between 22.1 and 46.1 microg dissolved Zn/L. This resulted in deterministic risk characterization ratios (RCR) that were below 1 in all river basins, suggesting that there is no deterministic regional risk associated with current use patterns of Zn in these river basins. With the probabilistic approach we identified rather limited risks, i.e., between <0.4 and 18.3%. When the EU RAR approach was applied to the same monitoring datasets, deterministic risks were found in different river basins. A detailed analysis showed that this different deterministic conclusion of risk is mainly due to the fact that the EU RAR (i) uses an additional assessment factor of 2 to derive the PNEC and (ii) uses a more conservative approach for implementing bioavailability (BioF approach). We argue that the larger conservatism in the EU RAR mainly originates from decisions made to deal in a pragmatic way with (i) uncertainty related to the across-species extrapolation of BLMs and (ii) the relatively high sensitivity of some multi-species toxicity studies.

Fish consumption is a safe solution to increase the intake of long-chain n-3 fatty acids

OBJECTIVES Dietary intake of long-chain (LC) n-3 PUFA in developed countries is low compared with recommendations. Fish is naturally rich in LC n-3 PUFA, but is also a dietary source of heavy metals and organic pollutants. We investigated whether the recommendation for LC n-3 PUFA could be reached through fish consumption, without exceeding the provisional tolerable weekly intake of methylmercury (MeHg) and the tolerable weekly intake (TWI) of dioxin-like compounds. Also, the contribution of margarines enriched with LC n-3 PUFA was assessed. DESIGN Published nutrient and contaminant data were used in a probabilistic model to calculate the simultaneous nutrient and contaminant intake for different fish consumption scenarios. RESULTS The Belgian recommendation for EPA + DHA (0.3% of total energy intake) can be reached by consuming fatty fish a minimum of twice a week, or by varying between lean and fatty fish a minimum of three times a week. At this fish consumption level, MeHg intake is not an issue of toxicological concern. The intake of dioxin-like compounds approximates the TWI when consuming fatty fish more than twice a week, this being a potential toxicological risk because other food items also contribute to the weekly intake of dioxin-like compounds. Use of margarine enriched with LC n-3 PUFA can help to increase LC n-3 intake, on average by 159 mg/d. CONCLUSIONS Combination of regular fish consumption (twice a week) with important contribution of fatty fish species, in combination with regular consumption of margarine enriched with EPA + DHA, can be advised to achieve the recommendation for LC n-3 intake.

An emission time series generator for pollutant release modelling in urban areas

Dynamic priority pollutant (PP) fate models are being developed to assess appropriate strategies for limiting the release of PPs from urban sources and for treating PPs on a variety of spatial scales. Different possible sources of PP releases were mapped and both their release pattern and release factors were quantified as detailed as possible. This paper focuses on the link between the gathered PP sources data and the dynamic models of the urban environment. This link consists of (1) a method for the quantitative and structured storage of temporal emission pattern information, (2) the retrieval of spatial emission source data from a GIS covering the studied urban area, (3) the coupling of the (GIS-based) spatial emission source data with temporal emission pattern information and (4) the generation of PP release time series to feed the dynamic sewer catchment model. Steps 3 and 4 were included as the main features of a dedicated software tool. Finally, this paper also illustrates the methods applicability to generate model input time series for generic pollutants (N, P and COD/BOD) in addition to priority pollutants.

Probabilistic environmental risk assessment of zinc in dutch surface waters

In the framework of the European Union (EU) New and Existing Chemicals Policy, a regional risk assessment for Zn according to the current technical guidance documents and a probabilistic approach, by mathematically integrating both best-fitting exposure concentrations and species-sensitivity distributions into a probabilistic risk quotient distribution using Monte Carlo analysis, was explored for The Netherlands. Zinc is an essential element, and the current probability distributions may not adequately deal with this property. The threshold Pareto distribution provided the best fit to the chronic Zn toxicity data, resulting in a predicted-no-effect concentration (PNECadd) for dissolved Zn of 34.2 microg/L, whereas use of the conventional normal distribution resulted in a PNECadd for dissolved Zn of 14.6 microg/L. The extracted exposure data resulted in a regional predicted environmental concentration (PEC) for dissolved Zn in the Dutch surface waters of 20.1 microg/L and in PECadd values for dissolved Zn of between 15.5 and 17.3 microg/L, depending on the background correction used. The conventional deterministic risk characterization identified a regional risk for Zn in the Dutch surface waters. The more comprehensive probabilistic approach used in the present study, however, identified only very limited potential risks for the Dutch region. A probabilistic median risk, that the environmental concentration is greater than the no-observed-effect concentration of a species in Dutch surface waters (0.5-0.6%), depending on the inclusion of background correction, was obtained from the best-fitting distributions. Because probabilistic approaches provide a quantifiable and improved assessment of risk and quantification of the uncertainty associated with that assessment, these techniques may be considered as a way to improve the EU risk assessment procedures for data-rich substances.

Determining environmental standards using bootstrapping, bayesian and maximum likelihood techniques: a comparative study

Environmental standards must be set in ways which give full recognition to all sources of uncertainty and variability of the toxicity data used to derive these standards. Toxicity data such as NOECs form a variability distribution describing species sensitivity distribution (SSD). In EU environmental regulations the 5th-percentile of SSD is used to set the quality criteria. In this paper, a comparison is made between the application of techniques characterising uncertainty and variability (bootstrap, maximum likelihood estimation (MLE) and Bayesian approaches) using small toxicity data sets to calculate the 5th-percentile. Estimating lower and upper uncertainty bounds of a specific percentile gives different results when different methods are used. Bayesian and MLE methods were found to be superior to parametric bootstrapping because they are easier to use and not so computationally intensive.

How to Use Secondary Data on Seafood Contamination for Probabilistic Exposure Assessment Purposes? Main Problems and Potential Solutions

ABSTRACT Seafood consumption is related to both favorable health benefits of nutrients and to potential adverse health impacts of chemical contamination. To quantify the magnitude of this dilemma, probabilistic intake assessments can be performed. One step in such a procedure involves the determination of nutrient and contaminant concentrations in seafood for which data need to be collected. This article describes the process of building up a database containing previously published contaminant concentrations in seafood, and defining input distributions characterizing the variability. During the constitution of the database, several problems influencing the comparability of the data were encountered related to (1) sampling plans of the published data; (2) sample handling prior to analysis; (3) analytical methodologies; (4) the format of reporting results; and (5) missing data. Different solutions for these methodological problems have been developed. Contaminant concentrations ranges per gram fresh weight of 2.4–4390.0 ng for mercury, 0.1–5736.6 ng for the sum of indicator PCB, 0.002–115.000 pg TEQ for the sum of all PCBs, 0.002–34.400 pg TEQ for dioxins, and 0.006–126.000 pg TEQ for total of dioxin-like compounds were found. This work confirms that more guidelines are needed to standardize the analytical methodologies to be used and the format for result reporting in order to improve the comparability of data critical to conduct a human intake and risk-benefit assessment.

Limitations of current risk characterization methods in probabilistic environmental risk assessment

In probabilistic environmental risk assessment, the likelihood and the extent of adverse effects occurring in ecological systems because of exposure(s) to substances are estimated. It is based on the comparison of an exposure/environmental concentration distribution, with a species sensitivity distribution derived from toxicity data. The calculation of a probabilistic risk can be performed in many ways (e.g., area under the curve in joint probability curves). However, several (hypothetical) examples and some theoretical considerations illustrate that the current risk characterisation methods have an integrative character and they focus on the statistical comparison of two distributions without properly considering the environmental interpretation of these underlying distributions. Several scenarios with varying exposure/environmental concentration distribution and species sensitivity distribution standard deviations are discussed.

Probabilistic intake assessment of polybrominated diphenyl ethers and omega-3 fatty acids through fish consumption

Food intake is one of the principal exposure routes of polybrominated diphenyl ethers (PBDEs) in humans. This study focuses on fish consumption as a PBDE exposure route. A probabilistic intake assessment of PBDEs and healthy long chain omega-3 PUFAs (LC n-3 PUFAs) was conducted for Belgian fish consumers in order to study the balance of the intake of LC n-3 PUFAs and PBDEs. Based on the observed fish consumption level in the sample, the mean intake of brominated diphenyl ether (BDE)-28, 47, 99, 100, 153, and 154 via fish was 0.85 ng/kg body weight (bw)/day and the intake of LC n-3 PUFAs was 3.45 mg/kg bw/day, being low compared to the recommendations. Scenario analyses showed that consuming 150 g salmon twice a week is advisable to achieve the recommended LC n-3 PUFA intake with a rather low PBDE intake. When replacing 150 g salmon by herring, the PBDE intake is higher without an increase in LC n-3 PUFAs. In contrast, the combination of cod and salmon leads to a similar PBDE intake compared to twice a week salmon, but to a lower LC n-3 PUFA intake. In conclusion, the methodology presented in the paper allows balancing benefits and risks related to fish consumption.

IMPROVING SEDIMENT-QUALITY GUIDELINES FOR NICKEL: DEVELOPMENT AND APPLICATION OF PREDICTIVE BIOAVAILABILITY MODELS TO ASSESS CHRONIC TOXICITY OF NICKEL IN FRESHWATER SEDIMENTS

Within the framework of European Union chemical legislations an extensive data set on the chronic toxicity of sediment nickel has been generated. In the initial phase of testing, tests were conducted with 8 taxa of benthic invertebrates in 2 nickel-spiked sediments, including 1 reasonable worst-case sediment with low concentrations of acid-volatile sulfide (AVS) and total organic carbon. The following species were tested: amphipods (Hyalella azteca, Gammarus pseudolimnaeus), mayflies (Hexagenia sp.), oligochaetes (Tubifex tubifex, Lumbriculus variegatus), mussels (Lampsilis siliquoidea), and midges (Chironomus dilutus, Chironomus riparius). In the second phase, tests were conducted with the most sensitive species in 6 additional spiked sediments, thus generating chronic toxicity data for a total of 8 nickel-spiked sediments. A species sensitivity distribution was elaborated based on 10% effective concentrations yielding a threshold value of 94 mg Ni/kg dry weight under reasonable worst-case conditions. Data from all sediments were used to model predictive bioavailability relationships between chronic toxicity thresholds (20% effective concentrations) and AVS and Fe, and these models were used to derive site-specific sediment-quality criteria. Normalization of toxicity values reduced the intersediment variability in toxicity values significantly for the amphipod species Hyalella azteca and G. pseudolimnaeus, but these relationships were less clearly defined for the mayfly Hexagenia sp. Application of the models to prevailing local conditions resulted in threshold values ranging from 126 mg to 281 mg Ni/kg dry weight, based on the AVS model, and 143 mg to 265 mg Ni/kg dry weight, based on the Fe model.

Evaluating the usefulness of dynamic pollutant fate models for implementing the EU Water Framework Directive

The European Water Framework Directive (WFD) aims at achieving a good ecological and chemical status of surface waters in river basins by 2015. The chemical status is considered good if the Environmental Quality Standards (EQSs) are met for all substances listed on the priority list and eight additional specific emerging substances. To check compliance with these standards, the WFD requires the establishment of monitoring programmes. The minimum measuring frequency for priority substances is currently set at once per month. This can result in non-representative sampling and increased probability of misinterpretation of the surface water quality status. To assist in the classification of the water body, the combined use of monitoring data and pollutant fate models is recommended. More specifically, dynamic models are suggested, as possible exceedance of the quality standards can be predicted by such models. In the presented work, four realistic scenarios are designed and discussed to illustrate the usefulness of dynamic pollutant fate models for implementing the WFD. They comprise a combination of two priority substances and two rivers, representative for Western Europe.