Regionalizing the supply chain in process life cycle inventory with multiregional input-output data: An implementation for ecoinvent with EXIOBASE

Abstract

Life cycle inventory (LCI) databases, such as ecoinvent, are crucial for life cycle assessment (LCA), but lack country-specific resolution in activity details and trade between regional activities in the supply chain. For example, ecoinvent only provides rest-of-world and European datasets for photovoltaic panel production, consuming the same global consumption mix of photovoltaic cells. Global or continental activities and trade limit the use of existing country activities, such as electricity production, and the accuracy and granularity of LCA results. We addressed this issue by disaggregating ecoinvent's global and continental unit processes to the country level and by regionalizing trade using country-sector-specific consumption mixes of product origins from the multiregional input-output table EXIOBASE. This produced a consistently regionalized ecoinvent at country level. Comparing the climate change impacts of 195,708 non-market datasets in our regionalized ecoinvent with reference datasets in ecoinvent reveals manufacturing sectors as the most affected. The study of photovoltaic panel production shows that the differences in climate change impacts are mainly due to the different country-specific suppliers of components and their electricity mixes, which are lost in the aggregated original data. The water consumption impacts of wheat production show the country differences only when regionalized CFs are applied, indicating that regionalized biosphere and technosphere flow amounts are needed for better results. The mapping quality between ecoinvent and EXIOBASE and inconsistent market definitions were discussed along with further limitations. Our study highlights the potential of incorporating better trade information and regional disaggregation in LCI databases to improve the LCA outcomes.