Investigation of carbon dioxide emission in China by primary component analysis

Abstract

Listen

Translate article

Principal component analysis (PCA) is employed to investigate the relationship between CO2 emissions (COEs) stemming from fossil fuel burning and cement manufacturing and their affecting factors. Eight affecting factors, namely, Population (P), Urban Population (UP); the Output Values of Primary Industry (PIOV), Secondary Industry (SIOV), and Tertiary Industry (TIOV); and the Proportions of Primary Industry's Output Value (PPIOV), Secondary Industry's Output Value (PSIOV), and Tertiary Industry's Output Value (PTIOV), are chosen. PCA is employed to eliminate the multicollinearity of the affecting factors. Two principal components, which can explain 92.86% of the variance of the eight affecting factors, are chosen as variables in the regression analysis. Ordinary least square regression is used to estimate multiple linear regression models, in which COEs and the principal components serve as dependent and independent variables, respectively. The results are given in the following. (1) Theoretically, the carbon intensities of PIOV, SIOV, and TIOV are 2573.4693, 552.7036, and 606.0791kt per one billion $, respectively. The incomplete statistical data, the different statistical standards, and the ideology of self sufficiency and peasantry appear to show that the carbon intensity of PIOV is higher than those of SIOV and TIOV in China. (2) PPIOV, PSIOV, and PTIOV influence the fluctuations of COE. The parameters of PPIOV, PSIOV, and PTIOV are −2706946.7564, 2557300.5450, and 3924767.9807kt, respectively. As the economic structure of China is strongly tied to technology level, the period when PIOV plays the leading position is characterized by lagging technology and economic developing. Thus, the influence of PPIOV has a negative value. As the increase of PSIOV and PTIOV is always followed by technological innovation and economic development, PSIOV and PTIOV have the opposite influence. (3) The carbon intensities of P and UP are 1.1029 and 1.7862kt per thousand people, respectively. The carbon intensity of the rural population can be inferred to be lower than 1.1029kt per thousand people. The characteristics of poverty and the use of bio-energy in rural areas result in a carbon intensity of the rural population that is lower than that of P.