Examining global biodiversity accounts: Implications of aggregating characterization factors from elementary flows in multi-regional input–output analysis

Extending multi-regional input–output (MRIO) models with spatially explicit life cycle impact assessment (LCIA) models allows practitioners to quantify biodiversity impacts at every step of global supply chains. Inconsistencies may be introduced, however, when high-resolution characterization factors (CFs) are aggregated so as to match the low spatial granularity of MRIO models. These aggregation errors are greater when CFs are aggregated via proxies, such as ecoregion land shares, instead of based on spatially explicit elementary stressor flows. Here, we describe our approach to tailoring application-specific CFs for use in MRIO studies. We apply a global agricultural production model, Spatial Production Allocation Model (MapSPAM), with the LCIA database, LC-IMPACT, to create crop-specific national CFs. We investigated i) if the differing aggregation approaches and the increased spatial explicitness of the constructed CFs deviate substantially from those in LC-IMPACT, and ii) what the resulting consequences for national production and consumption-based biodiversity footprints are when combining the tailor-made CFs with the EXIOBASE MRIO model. For the year 2020, we observe an increase in global production-based biodiversity impacts of 23.5% for land use when employing crop-specific CFs.