Abstract:In recent years, greenhouse gas (GHG) emissions have become a global focal point, and carbon emission accounting approaches have become even more important in measuring the reliability of data. This paper comprehensively analyzes the principles, reliability and other characteristics of the existing traditional carbon emission accounting approaches and comprehensively compares the application status and realizability of the carbon emission factor approach, the actual measurement approach and the mass balance approach. Additionally, this paper introduces the principle of using the relatively novel satellite remote sensing carbon emission accounting approach and the current status of its application, and proposes a CNN-LSTM model for carbon emission prediction based on electric power big data. Moreover, a kind of ARDL model is constructed for the data of Chinas Industrial Statistical Yearbook to explore the relationship between total carbon emissions and energy intensity, economic development level, and industrial development level, proving that there is a long-term linear relationship between them. It is concluded that a 1% increase in the level of economic development increases total carbon emissions by about 0.65%1.31%; a 1% increase in energy intensity increases total carbon emissions by about 1.23%1.95%; and a 1% increase in industrial development level increases total carbon emissions by about 0.27%0.43%.

Carbon emission accounting approaches study