Multi-region Input-Output-based Carbon and Energy Footprint Analysis of U.S. Manufacturing

Abstract

In this research, U.S. manufacturing activities' life cycle-based carbon and energy footprint impacts have been quantified, taking international trade linkages with the rest of the world into account. The U.S economy has been integrated into a multi-region input-output (MRIO) life cycle assessment framework where total of 40 major economies, including the USA, China, Russia, and others, plus the rest of the world (ROW) were modelled to assess global energy and carbon footprint impacts. Each country's economy is assumed to compromise 35 major industries based on the WIOD database classification. A total of 1435 (41 × 35 = 1435) industries has therefore been taken to represent the global structure of the world economy. The novelty of the approach is that the MRIO model has been developed in a stochastic fashion, plus global trade-linked uncertainties have also been taken into consideration. Top carbon emitting and energy consumer industries and countries have been analysed using data analytics and statistical modelling methods. The results show that the USA is the largest contributor to the total carbon footprint (CFP) and the total energy footprint (EFP) with 81.73% and 84%, respectively. Moreover, the agriculture/hunting forestry/fishing sector and the electricity/gas/water supply sectors dominate the overall U.S. carbon footprint, contributing 22% and 21.28%, respectively. The coke/refined petroleum/nuclear fuel sector has the largest share of the total energy footprint, with 47.9% of the total impacts.

Estimating the uncertainty of the greenhouse gas emission accounts in global multi-regional input–output analysis

“…A large part of environmentally extended GMRIO studies make only, if at all, qualitative considerations of uncertainty (Zhang et al, 2019). Studies that include quantitative considerations of parametric uncertainties include Lenzen et al (2010), Wilting (2012), Karstensen et al (2015), Moran et al (2018), Shrestha and Sun (2019), Zhang et al (2019), Kanemoto et al (2020) and Abbood et al (2023). Moreover, with Eora (Lenzen et al, 2013) and GLORIA (Lenzen et al, 2022), there exist two GMRIO databases that publish uncertainty estimates in the form of standard deviations alongside each data entry, thus allowing GMRIO practitioners to conduct an uncertainty analysis of their results by themselves.…”

“…In the case of the former, probability distributions are directly assigned to the input-output coefficients, i.e. the individual elements of the inter-industry, final-demand or satellite extension matrices (Abbood et al, 2023;Shrestha and Sun, 2019;Kanemoto et al, 2020). In the case of the latter, probability distributions are assigned to the raw data used to determine those coefficients (Lenzen et al, 2010(Lenzen et al, , 2013(Lenzen et al, , 2022Karstensen et al, 2015;Zhang et al, 2019).…”

“…In the first approach, simple heuristics are employed to assess the uncertainty of the raw data. This method is adopted by Wilting (2012), Abbood et al (2023), Kanemoto et al (2020) and Shrestha and Sun (2019).…”

Estimating the uncertainty of the greenhouse gas emission accounts in Global Multi-Regional Input-Output analysis

“…In case of the former, probability distributions are directly assigned to the input-output coefficients, i.e. the individual elements of the inter-industry, final demand, or satellite extension matrices (Abbood et al, 2023;Shrestha and Sun, 2019;Kanemoto et al, 2020). In case of the latter, probability distributions are assigned to the raw data used to determine those coefficients (Lenzen et al, 2010;Karstensen et al, 2015;Zhang et al, 2019;Lenzen et al, 2013Lenzen et al, , 2022.…”

“…In the first approach, simple heuristics are employed to assess the uncertainty of the raw data. This method is adopted by Wilting (2012); Abbood et al (2023); Kanemoto et al (2020); Shrestha and Sun (2019). For instance, Wilting (2012) applies heuristics such as: domestic input data being less uncertain than trade data, and sectors from one group being less uncertain than those from another, based on "differences in input characteristics."…”