An input-output-based Bayesian neural network method for analyzing carbon reduction potential: A case study of Guangdong province

Abstract

Economic development, population growth, industrialization and urbanization have led to large increases in anthropogenic carbon emission that has caused a variety of negative impacts on climate change and eco-environment systems. This study develops an input-output-based Bayesian neural network (IO-BNN) method for simulating the carbon emission of various economic sectors and generating the desired schemes of emission reduction. IO-BNN is applied to Guangdong province to identify its carbon emission path, carbon peak, and carbon reduction potential over a long-term planning horizon (2021–2050), in which multiple scenarios are designed to examine the effects of different environmental policies on economic and energy activities. Major findings are: (i) the key sectors and factors affecting Guangdong's carbon emission and economic development are equipment (Equ), construction (Con), transport and storage (Tra), other service (Oth), per capita energy consumption (CEC), and primary energy consumption (EC); (ii) under different environmental policies, Guangdong's carbon emission would reach the peak during 2025–2035 and then continuously decrease during 2036–2050; (iii) Guangdong's carbon emission would peak in 2025 under the optimal scenario, associated with adjustment of industrial structure (i.e. part of secondary industry would be shifted to tertiary industry) as well as reduction of primary energy consumption.