Determinants investigation and peak prediction of CO2 emissions in China’s transport sector utilizing bio-inspired extreme learning machine

• Published in: Environmental science and pollution research international 2021-10, Vol.28 (39), p.55535-55553
• Main Authors: Wang, Weijun, Wang, Jixian
• Format: Article
• Language: English
• Subjects: Aquatic Pollution; Artificial neural networks; Atmospheric Protection/Air Quality Control/Air Pollution; Biomimetics; Carbon; Carbon dioxide; Carbon dioxide emissions; Earth and Environmental Science; Ecotoxicology; Electrification; Emissions; Emitters; Environment; Environmental Chemistry; Environmental Health; Environmental science; Fossil fuels; Learning algorithms; Machine learning; Mathematical models; Optimization; Paris Agreement; Pollution abatement; Prediction models; Research Article; Sustainable development; Transportation industry; Urbanization; Waste Water Technology; Water Management; Water Pollution Control
• ISSN: 0944-1344; 1614-7499
• DOI: 10.1007/s11356-021-14852-z

The transport sector is recognized as one of the largest carbon emitters. To achieve China’s carbon peak commitment in the Paris Agreement on schedule, it is indispensable to explore the peak carbon emissions and mitigation strategies in the transport sector. Many researches in the past have contextualized in China’s total emissions peak, while the study about forecasting China’s transport CO 2 emissions peak seldom appeared, especially the application of intelligent prediction model. To further investigate the determinants and forecast the peak of transport CO 2 emissions in China accurately, a novel bio-inspired prediction model is proposed in this paper, namely, the extreme learning machine (ELM) optimized by manta rays foraging optimization (MRFO), hereafter referred as MRFO-ELM. Adhering to this hybrid model, the mean impact value (MIV) method is then employed to evaluate and differentiate the importance of thirteen influencing factors. Additionally, three scenarios are set to conduct prediction of China’s transport CO 2 emissions. The empirical results indicate that the proposed MRFO-ELM has excellent performance in terms of the optimization searching velocity and prediction accuracy. Simultaneously the level of vehicle electrification is verified to be one of the emerging major factors affecting China’s transport CO 2 emissions. The transport CO 2 emissions in China would peak in 2039 under the baseline model scenario, while the plateau would occur in 2035 or 2043 under sustainable development mode and high growth mode, respectively. The peak years imply much pressure on China’s transport carbon emissions abatement currently, whereas active policy adjustments can effectively urge the earlier occurrence of the emission peak. These new findings suggest that it is essential for China to improve the energy mix and encourage the electric energy replacement in line with urbanization pace, so as to achieve CO 2 emissions mitigation in the transport industry.