Evaluation of LCIA characterization models for marine ecotoxicity

The marine ecotoxicity impact category is becoming increasingly important given the threats faced by marine ecosystems. This category still does not have a characterization model consensually accepted by the scientific community, which poses a challenge for LCA practitioners in selecting which method to apply. To date, this category has been assessed through general criteria that do not always consider the specificities of this environment. This study develops specific criteria to evaluate marine ecotoxicity characterization models.

The development of this research proceeded through the following steps: (a) identification of marine ecotoxicity characterization models through a systematic literature review; (b) formulation of exclusionary criteria to select models from the identified pool with sufficient applicability and operability to get included in the subsequent assessment; (c) development of evaluative criteria based on previous methodologies, enabling a comprehensive assessment of the selected models in terms of their ability to adequately represent ecotoxicological impacts on the marine environment and also the needs of the oil and gas industry sector; (d) application of the established criteria to the pre-selected models; and (e) application of the three best-ranked models in a case study for the oil industry, to illustrate and discuss differences in the obtained results.

The application of four evaluative criteria to the six previously identified and selected characterization models revealed significant variability among them, concerning their spatial comprehensiveness, elementary flows, environmental relevance, and documentation quality. The best average performance of the models was in Criterion 3—Scientific Robustness and Transparency, while the worst performance was in Criterion 4, developed specifically to verify the suitability of the models for the oil sector. The model by Verones et al. (2020) demonstrated the best overall performance, followed by the models by Dong et al. (2016; 2018) and USES-LCA 2.0. Consensus results in relative terms were not reached when applying these models in the case study, highlighting the influence that the choice of the characterization model has on the final results.

As the marine ecotoxicity impact category lacks a universally accepted characterization model, the application of at least two models and the use of their results in a complementary manner appears to be the most appropriate approach. Although all these models have limitations and associated uncertainties, their application is preferable to omitting results for the category. The selection presented should not be understood as a general rule, but rather as an alternative for guidance.