Exploring the Influence of Database Selection on the Life Cycle Assessment of Digital Services

Wattiez, Augustin [UCL] Kelly Le Goff [Mavana] Bol, David [UCL]

In Life Cycle Assessment (LCA) studies of digital services, practitioners rely on databases that model electronic components, devices, and systems. However, it is acknowledged in the scientific literature that LCA results can be sensitive to the choice of database. For this reason, the present work aims to assess how database selection influences (1) absolute impact estimates and (2) contribution analysis results when conducting a LCA for a digital service. We modeled a case study - music streaming - using three databases: Managed LCA Content, CODDE, and Ecoinvent. This work highlights significant variations in impact estimates for specific indicators. Additionally, environmental hotspots may differ when using different databases. These disparities stem from inconsistencies in background modeling and in granularity between the databases, especially for integrated circuits and Li-ion batteries.