Data quality assessment of aggregated LCI datasets: A case study on fossil‐based and bio‐based plastic food packaging

• Published in: Journal of industrial ecology 2024-12, Vol.28 (6), p.1900-1911
• Main Authors: Carlesso, Anna, Pizzol, Lisa, Marcomini, Antonio, Semenzin, Elena
• Format: Article
• Language: English
• Subjects: aggregated LCI datasets; Application; Data quality; data quality assessment; Datasets; Environmental impact; Evaluation; Food packaging; Food quality; Fossils; industrial ecology; Information processing; Life cycle analysis; Life cycle assessment; life cycle inventory; Life cycles; Metadata; Packaging; plastic food packaging; Plastics; Polymers; Quality assessment; Quality control; Quality management; Reliability; Representativeness; Transparency
• ISSN: 1088-1980; 1530-9290
• DOI: 10.1111/jiec.13572

Environmental impacts resulting from plastic food packaging, made from both fossil‐based and bio‐based polymers, are increasingly analyzed in life cycle assessment (LCA) studies. However, the literature reveals significant variations in results for the same polymer within the same scope. To enhance the reliability of these assessments, data quality assessment (DQA) plays a relevant role. However, despite most of the LCA studies employing aggregated life cycle inventory (LCI) datasets, in the literature, DQA methods for aggregated processes are not available. To fill this gap, in this paper, a DQA for aggregated LCI datasets is proposed and demonstrated through its application to 101 aggregated LCI datasets, extracted from Ecoinvent and GaBi databases. The DQA method has been developed by adapting and integrating the pedigree matrix and the data quality ranking proposed by the recently published EC Plastic LCA method. The three data quality indicators (DQIs) used are technological, geographical, and time‐related representativeness. The application of this method exhibits an overall positive evaluation of the selected datasets with differences among the three DQIs. Moreover, it highlights the role of metadata structure in adequately supporting a robust DQA. Indeed, in the absence of a common framework that defines, assesses, and provides access to data quality information, transparency must be assured by the operator in the metadata interpretation and related assumptions along the DQA process. Finally, although the proposed DQA method was developed for the plastic sector, its application can be extended to LCI aggregated datasets relevant to other sectors, materials, and products.