Background. There is an increasing interest to obtain holistic assessments of environmental impact of chemicals, i.e. integrating aspects such as aquatic ecotoxicity, eutrophication, climate change etc. To that end, life cycle assessment (LCA) is seen as key assessment method. The currently ongoing ‘Product Environmental Footprint’ (PEF) project by the European Commission collects experience in the application of the LCA methodology to the assessment of products. The ultimate objective of this project is to provide LCA-based metrics as a basis for distinguishing between products according to their overall environmental impact. Aquatic ecotoxicity is one of the impacts assessed within LCA. At the same time, there is the understanding that the currently used methodology for environmental risk assessments (ERA) of chemicals is limited with regard to characterising environmental impact of chemicals.
Objectives. The assessment of the risk of adverse effects in the environment originating from the uses and emissions of chemicals is at the heart of ECETOC’s activities. Hence, ECETOC established a task force to evaluate the extent to which the limitations of the environmental risk assessment methodology are relevant for the ecotoxicity assessment in LCA, and to identify options for including the ecotoxicity impact of chemicals into assessments of products reflecting their environmental impact in a holistic manner. The following terms of reference guided the work of the task force.
Terms of Reference
- Conduct a scientific evaluation of the USEtox method, currently the leading method for calculating ecotoxicity impacts in LCA, and its relevance to the real world;
- Investigate the relationship of USEtox to chemical risk assessment methodology;
- Provide guidance on the scientific relevance and interpretation of USEtox results in the context of chemical impact assessment and selection of chemical-based (manufactured) products.
The task force started out with a comparison of the ecological risk assessment versus LCA as predictive approaches for approximating environmental impact of chemicals (Chapter 2). With regard to LCA the focus was put on the ecotoxicity evaluation via the USEtox methodology. This comparison was complemented by another comparison between the USEtox methodology and the ‘Critical Dilution Volume’-(CDV) method (Chapter 3). This method is applied for assessing impact of chemical-based products, e.g. in the EU EcoLabel schemes for detergent and personal care products. In addition, Chapter 3 also provides brief outlines of the ‘Environmental Safety Check’ and the ‘ProScale’-method. Beyond the conceptual aspects, this comparison included a case study (Chapters 4 and 5), which highlighted the differences between LCA and CDV regarding the assessment methodology and issues regarding availability and quality of data.
Building on the results of Chapter 2, the USEtox methodology for assessing ecotoxicity is put into the perspective of the greenhouse gas impact assessment in LCA (Chapter 6). This chapter also contains a reflection on the joint report ‘New Challenges in Risk Assessment’ by the EU Commission’s scientific committees on newly identified health risks, on health and environmental risks and on consumer protection. Finally, Chapter 7 outlines options for proceeding with ecotoxicity in holistic environmental product impact assessment, taking into account the results of the preceding Chapters.
Outcome – Comparison of LCA versus Ecological Risk Assessment
The comparison of the ecological risk assessment versus the freshwater ecotoxicity assessment concluded that both methods (and CDV and ProScale as well) are based on the same type of input information. In addition, the metric is obtained in a very similar manner, i.e. by using single-species ecotoxicity data as a reference point for deriving the impact metric. Similarly, in lower-tier risk assessment and in LCA the exposure assessment assumes that the concentrations do not vary in time and space. Hence, there are large methodological similarities.
However, there are discrepancies regarding the assessment objectives, the scopes of the assessment, and the efficiency strategies applied in the assessment. Ecological risk assessment attempts to demonstrate that ecological impact is unlikely and that no risk management is needed. From that point of view, these assessments typically start with a low initial demand for accuracy and proceed in a tiered fashion. They often start with conservative / worst-case assumptions to avoid false negative conclusions regarding the safe use of a chemical. If needed, a stepwise increase of realism is accomplished by refining the input data or turning to refined / higher tier prediction tools. In contrast, LCA attempts to provide a numerical expression of the ecotoxicity impact. Hence, the ambition is to be realistic. The ecotoxicity assessment in LCA addresses the emissions originating from the entire life cycle of a product or a service and hence addresses a multitude of chemicals. Chemical risk assessment typically considers the single use or a single life cycle step of a
The major efficiency strategy in risk assessment is to take a tiered approach to increasing the realism of the assessment. This allows to invest the minimum of effort needed for demonstrating that the predicted environmental concentrations do not exceed the effect thresholds. Hence, ecological risk assessments can have a very low to a very high degree of realism. In order to solve the practical difficulty to obtain the large amount of data required to assess all the emissions along the entire life-cycle, so-called life cycle inventories are an essential element for applying LCAs. The life cycle inventories contain the emissions of individual life cycle steps. The emissions along the entire life cycle are obtained as the sum of the emissions, which for many life cycle steps are obtained from the life cycle inventories.
We describe several mechanisms by which uncertainties are introduced into the impact assessment stage of an LCA. These can be due to uncertainties in the characterisation factors or missing characterisation factors. Additionally, differences in granularity and perspective of the datasets in LCI databases can give rise to apparent differences in impact size. However, these data gaps and differences in granularity are not obvious and can at present only be found by tedious analysis.
Outcome – Case studies
In the case study a virtual product consisting of several chemicals is assessed. Disposal of the product via wastewater is part of the life cycle. This disposal scenario represents a worst-case for aquatic ecotoxicity and can also be covered by the CDV method. According to the LCA assessment results obtained with USEtox, impact of the disposal of the product to the wastewater is at least orders of magnitude larger than that of the earlier life cycle stages. This indicates that the earlier life cycle stages may be negligible for assessing the impact of products which are disposed of via wastewater. The results of the CDV assessments are poorly correlated to those of the assessment with USEtox with regard to the ranking and the absolute scores of the constituents of the product in the scope of our case study. Differences in results obtained with USEtox
and CDV are based on differences in the models, but are also due to the data used in the parametrization of either model.
For assessment of the impact of the disposal stage in LCA with USEtox and for the CDV assessment, the input data set is amended in order to obtain a complete set of characterisation factors. These summarise the ecotoxicological properties such as degradability, partitioning behaviour and ecotoxicity. The LCA of the earlier life-cycle stages has to rely on the database data. Closer analysis of these data reveals that characterisation factors for potentially relevant substances are missing. This lack of data is not easily visible in current LCA practice and requires a tedious manual analysis of the results. As a consequence, the LCA assessor cannot know the quality of the LCA result without checking the quality of the database data.
Outcome – Putting LCA into perspective
The comparison of the greenhouse gas assessment with the ecotoxicity assessment in LCA indicates that the latter is much more complex and suggests that the result of the ecotoxicity assessment is significantly less certain than that of the greenhouse gas assessment. The aspect of the uncertainty of the ecotoxicity assessment is followed up by considering the degree of realism of the ecological risk assessment. This is relevant given the methodological similarities between the ecotoxicity assessments in LCA and ecological risk assessment given that the LCA ecotoxicology assessment bears much resemblance to an initial tier environmental risk assessment of a mixture of chemicals. According to the scientific committees of the EU commission (Scientific Committee on Consumer Safety – SCENIHR; Scientific Committee on Health and Environmental Risks – SCHER; Scientific Committee on Consumer Safety – SCCS), the methodology underlying ecological risk assessment is sufficiently conservative for identifying ecological risks. However, it is too uncertain to provide a realistic characterisation of the impact of chemicals on the aquatic environment. In summary, it is concluded that the ecotoxicological assessment methodology is unsuitable for characterising the actual environmental impact because of large uncertainties in results. In consequence, these may not be a valid basis for distinguishing products from each other based on the real environmental impact.
Addressing Ecotoxicity for Comparing Products / Services – Possible ways forward
Nevertheless, there is a considerable demand for comprehensive evaluations of the environmental impact of products and services. Hence, the task force developed options for addressing this need in the future. These options are intended as a basis for a discussion for a multi-stakeholder discussion, e.g. in a joint workshop of the EU Commission, independent experts, and participants of the EU product environmental footprint project.
Within these options, there are two options addressing ecotoxicity without the result being reflected in the LCA assessment. Four additional options consider ecotoxicity as an integral part of life cycle assessment. The most obvious (and the most difficult to implement) option is to improve the ecological risk assessment methodology and subsequently the LCA methodology that would be based on it. The target is to achieve a degree of certainty enabling the size of the ecotoxicological impact of one product to be compared to that of another product. A second option is to make use of the abundance of REACH data in the current USEtox methodology in order to increase the number of available characterisation factors for chemicals and their quality. This will primarily improve the comparability of LCA-ecotoxicity results between products. Option three builds on this and aims at harmonising the use of ecotoxicity data between LCA and ecological risk assessment. It consists of changing the ecotoxicity reference data from the geometric mean of the EC50-values (current in USEtox) to the PNEC (Predicted no effect concentration), which protect the most sensitive species. The last option is not to change the assessment and to obtain an understanding of the discriminative power of the current ecotoxicity assessment in USEtox.