基于2010—2023年中国30个省（自治区、直辖市）面板数据，采用IPCC投入产出法系统核算城市交通碳排放量，并运用XGBoost-SHAP可解释机器学习模型分析六大驱动因素的非线性影响。研究结果表明：（1）城市交通碳排放呈现显著的区域差异与阶段性演化特征；（2）各驱动因素对交通碳排放的影响普遍存在非线性特征与阈值效应；（3）VAT与FTK是推动碳排放增长的核心因素，PC是抑制碳排放的关键变量，PGDP呈“倒U型”关系，ET呈正向边际递减特征，RPK表现为多阶段波动效应。研究可为精准制定城市交通领域低碳政策提供科学支撑。

城市交通碳排放；非线性影响；XGBoost-SHAP；低碳转型

[1]IPCC.ClimateChange2023:SynthesisReport[R].Geneva:IPCC,2023.

[2]IEA.GlobalEVOutlook2023[R].Paris:InternationalEnergyAgency,2023.

[3]WangH,OuX,ZhangX.Mode,technology,energyconsumption,andresultingCO2emissionsinChina'stransportsectorupto2050[J].EnergyPolicy,2017,109:719-733.

[4]ZhangX,LiY,WangS,etal.AnanalysisofthedecompositionanddrivingforceofcarbonemissionsintransportsectorinChina[J].ScientificReports,2024,14:28743.

[5]ChengZ,LiL,LiuJ.Industrialstructure,technicalprogressandcarbonintensityinChina'sprovinces[J].RenewableandSustainableEnergyReviews,2018,81:2935-2946.

[6]WangY,LiaoM,etal.Unearthingtheimpactofurbansprawlontransportcarbondioxideemissions:freshevidencefrom274Chinesecities[J].JournalofGeographicalSciences,2024,34(11):2173-2196.

[7]GrossmanGM,KruegerAB.Economicgrowthandtheenvironment[J].QuarterlyJournalofEconomics,1995,110(2):353-377.

[8]ChenT,GuestrinC.XGBoost:Ascalabletreeboostingsystem[C].KDD2016:785-794.

[9]XingM,WangZ,etal.AnXGBoost-SHAPanalysisofthedrivingfactorsofcarbonemissionsinChina'sfirst-tiercities[J].ScientificReports,2025,15:8601.

[10]IPCC.2006IPCCGuidelinesforNationalGreenhouseGasInventories[R].Hayama:IGES,2006.