摘要
区域交通减排是实现全局减排和交通低碳发展的关键.为助力中国交通运输业提前实现碳达峰、碳中和战略发展目标,利用2000~2021年中国30个省(市、自治区)的面板数据,采用多种机器学习回归算法,构建了区域交通运输业碳排放预测模型,其中Lasso回归模型与支持向量机算法相结合所建立的预测模型表现最佳.以粤沪鲁川4省(市)的交通运输业为例,设定了基准、节能减排和技术减排这3种未来发展情景,并运用预测模型(Lasso_SVM)对粤沪鲁川2022~2035年交通运输碳排放量进行了预测.结果表明,在基准、节能减排和技术减排情景下,粤沪鲁川最早实现交通碳排放达峰时间节点分别为2029年、2028年、2030年和2029年,其峰值分别为73.59、52.16、55.08和33.46 Mt.最后,结合4省(市)在不同情景下的碳排放预测结果,制定了科学可行的减排路径,为促进中国交通运输业提前实现碳达峰提供技术支持.
Reduction in regional traffic emission is crucial for achieving overall emission reduction and low-carbon development in the transportation sector.To assist China's transportation sector in early realizing of its carbon peak and carbon neutrality goals,panel data from 30 provinces(municipalities,autonomous regions)of China from 2000 to 2021 were utilized.Various machine learning regression algorithms were employed to construct a predictive model for regional transportation carbon emissions,among which the model combining the Lasso regression and support vector machine algorithms performed the best.Taking the transportation sector in Guangdong,Shanghai,Shandong,and Sichuan as examples,three future development scenarios—baseline,energy-saving emission reduction,and technology-driven emission reduction—were set.The predictive model(Lasso_SVM)was used to forecast the carbon emissions from transport sector in these provinces from 2022 to 2035.The results indicated that under the baseline,energy-saving emission reduction,and technology-driven emission reduction scenarios,the earliest peak times for carbon emissions from transport sector in Guangdong,Shanghai,Shandong,and Sichuan were the years 2029,2028,2030,and 2029,respectively,with peak values of 73.59,52.16,55.08,and 33.46 Mt,respectively.Finally,based on the carbon emission forecasts under different scenarios for the four provinces,scientifically feasible emission reduction pathways were formulated to provide technical support for advancing the carbon peak achievement in China's transport sector.
作者
宋永朝
舒秦
金程容
郑少鹏
罗亮
SONG Yong-chao;SHU Qin;JIN Cheng-rong;ZHENG Shao-peng;LUO Liang(School of Traffic&Transportation,Chongqing Jiaotong University,Chongqing 400074,China;Broadvision Engineering Consultants Co.,Ltd.,Kunming 650220,China;Railway No.5 Bureau Group First Engineering Co.,Ltd.,Changsha 410000,China)
出处
《环境科学》
北大核心
2025年第4期1995-2008,共14页
Environmental Science
基金
云南省科技厅基础研究项目(202401AT070060)
云南交投集团科创项目(YCIC-YF-2021-09)。
关键词
交通运输
碳排放
机器学习
情景预测
减排路径
transportation
carbon emissions
machine learning
scenario prediction
emission reduction pathways
