中国交通碳排放及影响因素时空异质性

选取30个省级行政单位作为空间单元，采用探索性空间数据分析（ESDA）方法对交通碳排放时空分布格局进行研究，同时考虑空间单元的差异性，构建地理加权回归（GWR）模型分析交通碳排放影响因素的时空异质性.研究发现：2000~2015年交通碳排放量呈现显著的空间聚类特征，聚类趋势逐年加强.双变量空间自相关指数为0.165~0.274，显著性水平介于0.016~0.045，表明交通碳排放同机动车保有量、GDP、货运周转量及客运周转量之间存在显著的空间正相关关系.GWR模型的R²在0.783~0.865之间，而OLS模型的R²在0.675~0.844之间，且GWR模型的AICc值均低于OLS模型的，说明GWR模型的拟合结果明显优于OLS模型，可以更好地解释交通碳排放的影响机制.GWR的回归结果表明碳排放的影响因素存在明显的时空异质性特征，其中GDP是主要的推动因素，部分地区回归系数高达0.91，2000年影响程度由东向西递减，而2005、2010和2015年由北向南递减.客运周转量起到关键的抑制作用，影响程度由东北向西南递减.因此建议应当充分考虑碳排放影响因素的时空异质性特征，制定差异化的碳减排政策.

Spatio-temporal heterogeneity of transportation carbon emissions and its influencing factors in China

Taking the 30 provinces in Mainland China as spatial analysis unit, the exploratory spatial data analysis (ESDA) method was employed to explore the spatio-temporal pattern of the transportation carbon emissions. Moreover, considering the spatial non-stationary, the geographically weighted regression (GWR) model was applied to analyze the spatio-temporal heterogeneity in the influencing factors of the transportation carbon emissions. The results indicated a significant spatial agglomeration in the transportation carbon emissions, and showing a gradual upward trend across time during the studied period 2000~2015. The Moran's I indices of the Bivariate spatial autocorrelation were 0.165~0.274 and the statistical significance levels were 0.016~0.045, indicating that there was a significant spatial positive correlation between the transportation carbon emissions and the variables, such as motor vehicle population, GDP, freight turnover and passenger turnover. The R² of the GWR models were between 0.783 and 0.865, while the R² of the OLS models were between 0.675 and 0.844; moreover, the AICc values of the GWR model were lower than those of the OLS models', demonstrating the goodness of the GWR models compared to the OLS models. This indicates that we can use the outcomes of the GWR models to better explain the impact mechanism of the transportation carbon emissions. The analysis of the GWR revealed that the influencing factors of the transportation carbon emissions had obvious spatio-temporal heterogeneity. GDP was among the major driving factors, with regression coefficient as high as 0.91 in some areas. The impact of GDP decreased from east to west in 2000, while decreasing from north to south in 2005, 2010, and 2015. The passenger turnover played a key inhibitory role, with its influence decreasing from northeast to southwest in all of the study years. In this context, the spatio-temporal heterogeneity of carbon emission influencing factors should be fully understood to formulate differentiated carbon emission reduction policies.