Satellite account dependent sensitivity differences to source data uncertainty in environmentally extended multi-regional input-output footprint results

Environmentally extended multi-regional input-output (EE-MRIO) models play a crucial role in sustainability analysis and policy making, yet the treatment of uncertainty in MRIO modelling remains under-explored. So far, uncertainty analysis has primarily been applied to carbon footprints and methods such as aggregation-based uncertainty analyses and Monte Carlo simulations have been used. In this work, we apply linear error propagation to trace uncertainty from environmental and social extensions to final footprint estimates, with validation through Monte Carlo simulations. We also conduct a sensitivity analysis to identify which footprints are more influenced by extension uncertainty. Our findings suggest that footprints linked to extensions evenly distributed across economic sectors (ex: GHG emissions, nitrogen, and employment) exhibit lower relative standard deviations compared to those tied to sector-specific extensions, such as land use, phosphorous, and water consumption. Sector-specific footprints, in general, demonstrate greater sensitivity to stressor uncertainty. The research further introduces an Outsource Coefficient, revealing that regions with higher outsourcing levels in their consumption-based accounts are less impacted by uncertainty in the underlying extension data. Finally, we provide uncertainty estimates for all EXIOBASE footprints using a 2019 product-by-product EXIOBASE 3 model and assume a uniform relative standard deviation of 0.1 across all extensions.