The "top-down" method was used to measure the traffic carbon emissions from 1985 to 2016 in the Yangtze River Economic Belt and analyze its spatial pattern and temporal evolution characteristics. Considering the unexpected output, a three-stage DEA model was constructed to evaluate and compare the traffic carbon emission efficiency of the Yangtze River Economic Belt, excluding the influence of external environment variables and random errors. The study found that first, the total traffic carbon emissions in the Yangtze River Economic Belt showed a rising trend, among which the carbon emissions from petroleum energy consumption accounted for the largest proportion. Sichuan, Hubei, and Hunan and the Su-Zhe-Hu Region were the high-value areas of traffic carbon emissions in the upper, middle, and lower reaches of the Yangtze River, respectively. Second, from the east to west, the center of traffic carbon emissions generally showed a changing track of moving east first and then west; from the north to south, it highlighted the characteristics of increasing concentrated distribution along the Yangtze River in space. Third, there was an obvious spatial differentiation in the traffic carbon emission efficiency values of different provinces; from 2007 to 2016, the efficiency value of the eastern region was the highest, but the value of the central region changed from higher than that in the western region to lower than that in the western region. Finally, external environmental factors had a significant impact on the efficiency of traffic carbon emissions, in which the optimization of industrial structure was found to be conducive to the improvement of traffic carbon emission efficiency, while the influence of government intervention was changed from "innovation compensation" effect to "compliance cost" effect.
Keywords: Yangtze River Economic Belt; efficiency analysis; spatial pattern; three-stage DEA; traffic carbon emissions.