Research on Spatial and Temporal Pattern Evolution and Driving Factors of County Carbon Emissions in Underdeveloped Regions: Gansu Province of Western China as an Example

Abstract

This paper used the Gini coefficient, standard deviation ellipse, and spatial autocorrelation model to analyze the overall changes, regional differences, spatio-temporal evolution pattern, and clustering characteristics of carbon emissions in 87 counties in Gansu Province from 1997 to 2017, based on which driving factors of carbon emissions were detected using the geographic detector model, so as to provide a reference for promoting low-carbon green development and ecological civilization construction in Gansu Province. The empirical research results found that county carbon emissions in Gansu Province showed a “first urgent and then slow” upward trend, and the difference in carbon emissions level has a slightly decreasing trend, and there are significant regional differences. Compared with other regions, the difference in county carbon emissions level in the Longzhong region has a smaller decline. Meanwhile, the county carbon emissions show spatial differentiation characteristics “medium-high and low-outside,” among which the carbon emissions in areas with better economic foundations are much higher than those in other areas, and the spatial polarization effect is obvious. In addition, there is a significant spatial positive correlation between county carbon emissions. The counties with high-high clusters are relatively stable, mainly concentrated in the Longzhong region, while counties with low-low clusters are slightly reduced, mainly concentrated in the southern ethnic region and the Longdongnan region, and the county carbon emission clusters type has a spatial locking effect. This is mainly due to the large differences in economic scale, industrial structure, and population size in Gansu Province, and the interaction between economic scale and other factors has a more significant impact on the spatial differentiation of carbon emissions. Moreover, the leading influencing factors of county carbon emission differences also have regional differences. Therefore, differentiated and targeted carbon emission reduction strategies need to be implemented urgently. Due to the lack of real county energy consumption statistics, the research results need to be further tested for robustness.