Carbon Emission Forecasting In China’s Transportation Industry Under Carbon Peaking Constraints

In 2020,President Xi Jinping solemnly announced that China will scale up its Intended Nationally Determined Contributions by adopting more vigorous policies and measures,aiming to have CO2 emissions peak before 2030.The transportation industry is one of the pillar sectors of China’s national economy,which is also under enormous pressure to reduce emissions.China Academy of Transportation Sciences points out that China’s transportation industry ranks third in carbon emissions after industry and construction,and will further increase,with greater upward pressure.At present,carbon emission control in the transportation industry is still unsatisfactory,so the forecast of future carbon emission in China’s transportation industry will help the government management to grasp the serious situation of carbon emission control in transportation,so as to accelerate the formulation of more stringent and scientific emission reduction policies.To this end,this paper firstly calculates the carbon emissions of the transportation sector in 30 provinces,cities and autonomous regions of China from 2000 to 2019 based on the "top-bottom" method proposed by IPCC,and establishes an extended STIRPAT model,selecting six influencing factors,namely GDP per capita,value added of transportation industry,energy intensity,clean energy share,transportation The six influencing factors of GDP per capita,value added of transportation,energy intensity,clean energy share,transportation turnover,and transportation intensity were selected as independent variables,and a ridge regression model was established using SPSS 26.0 to analyze the influencing factors of transportation carbon emissions.Secondly,to address the limitations of traditional prediction models in carbon emission prediction studies,and considering the small sample and nonlinear characteristics of carbon emission in transportation industry,In this paper,two machine learning models,BP neural network and support vector regression(SVR),are selected to simulate and compare the prediction accuracy of three models,including the traditional GM(1,n)prediction model,for the prediction of carbon emissions in China’s transportation industry.The results show that the machine learning model has a significantly better prediction effect,among which the support vector regression model has the smallest error index and is more suitable for the nonlinear and small sample prediction scenario in this paper.Meanwhile,in order to make up for the shortcomings of the parameter selection method of the support vector regression model,we introduce the Chicken Swarm Optimization and establish the CSO-SVR model in this paper.The error indexes indicated that,compared with the unoptimized model,the prediction accuracy of the optimized model has been significantly improved,and it has a strong global search capability,which can provide a better method to support the research of carbon emission prediction in the transportation industry.Finally,based on this model,the carbon emissions of China’s transportation industry are projected to peak in 2034,2035 and 2037 under the low-carbon,baseline and high-carbon scenarios,respectively.The comprehensive results of the above study will help provide data support and theoretical basis for the differentiated development of energy saving and emission reduction policies in China’s transportation industry,as well as provide reference for accelerating the achievement of the carbon peak target.