Nonlinear relationship between urban form and transport CO_(2)emissions:Evidence from Chinese cities based on machine learning  

基于机器学习的中国城市形态与交通二氧化碳排放之间的非线性关系研究

作　　者：LI Linna DENG Zilin HUANG Xiaoyan 李琳娜;邓梓林;黄晓燕

机构地区：[1]Faculty of Geographical Science,Beijing Normal University,Beijing 100875,China [2]Northwest Land and Resources Research Center,Global Regional and Urban Research Institute,Shaanxi Normal University,Xi'an 710119,China

出　　处：《Journal of Geographical Sciences》2024年第8期1558-1588,共31页地理学报（英文版）

基　　金：National Natural Science Foundation of China,No.42071227,No.42371214。

摘　　要：Reducing carbon emissions from the transport sector is essential for realizing the carbon neutrality goal in China.Despite substantial studies on the influence of urban form on transport cO_(2)emissions,most of them have treated the effects as a linear process,and few have studied their nonlinear relationships.This research focused on 274 Chinese cities in 2019 and applied the gradient-boosting decision tree(GBDT)model to investigate the nonlinear effects of four aspects of urban form,including compactness,complexity,scale,and fragmentation,on urban transport CO_(2)emissions.It was found that urban form contributed 20.48%to per capita transport CO_(2)emissions(PTCEs),which is less than the contribution of socioeconomic development but more than that of transport infrastructure.The contribution of urban form to total transport CO_(2)emissions(TCEs)was the lowest,at 14.3%.In particular,the effect of compactness on TCEs was negative within a threshold,while its effect on PTCEs showed an inverted U-shaped relationship.The effect of complexity on PTCEs was positive,and its effect on TCEs was nonlinear.The effect of scale on TCEs and PTCEs was positive within a threshold and negative beyond that threshold.The effect of fragmentation on TCEs was also nonlinear,while its effect on PTCEs was positively linear.These results show the complex effects of the urban form on transport CO_(2)emissions.Thus,strategies for optimizing urban form and reducing urban transport carbon emissions are recommended for the future.

关 键 词：urban form transport CO emissions nonlinear effect sustainable transport gradient-boosting decision treemodel 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程] X73[自动化与计算机技术—控制科学与工程]