Drivers of CO2 emissions from power generation in China based on modified structural decomposition analysis

Abstract

Currently, over 20% of global electricity and 30% of global CO2 emissions from fuel combustion are generated in China. Understanding the driving forces of CO2 emissions from power generation is critical for both decarbonizing the power sector and achieving national carbon reduction targets. The objectives of this study were to identify critical driving forces behind changes in CO2 emissions from the power sector and to propose appropriate decarbonization pathways. First, the generation and demand structures of the power sector were introduced into a CO2 emission accounting model and decomposition analysis. Instead of traditional input–output analysis, structural decomposition analysis modified using a power transmission table was used to investigate the impacts of five driving factors of CO2 emissions from China’s power generation. The five driving factors comprised the proportion of thermal power, power generation technology, power generation structure, power demand structure and power demand, whereby the latter was divided further into nine detailed parts. Considering five regional power grids in China, the contributions of these factors were analyzed at both national and regional level. The results showed that the majority of the increase in CO2 emissions during 2007–2012 could be attributed to electricity generation (96.2%) driven by changes in power demand, which should be the key to power sector decarbonization. By contrast, 30.7% of emissions were offset by changes in the proportion of thermal power and technology, demonstrating the obvious effects of China’s policy on clean energy transition. Additionally, all power grids exhibited an increase in CO2 emissions from electricity generation, with the east and central grids accounting for 64% of the national increase. Power transmission structure had only a small impact on CO2 emissions from power generation. By using the electricity transmission table, we modified SDA to overcome the time lag issues and eliminate the reliance on IO data, and continuous annual data rather than aggregated five-yearly data can be used to capture the structural effects, thus providing more precise results for the driving forces of emission changes. Our case study shows that there is huge potential in extending the IO-based SDA method to other trade-related studies.