Heterogeneity analysis on carbon emissions of region using geographically weighted regression

In this study, First, the 30 provinces (autonomous regions and municipalities) are selected as the basic space unit. Then, Geographically weighted regression (GWR) methods are employed to discover the factors and its spatial and temporal distribution for carbon emissions. Finally, the data from China Statistical Yearbook and China energy statistical yearbook from 2003, 2006 to 2010 is adopted to evaluate the reasonability of the proposed method. Our research findings are shown as follows: (1) The regions with a large amount of Carbon emissions are concentrated in mid-east region and its surrounding regions in central and eastern China between 2003 and 2010. (2) Impact factors of carbon emission have spatial temporal heterogeneity. For example, influence extent of GDP is diverse in different province and that the regression coefficients of GDP in 2006 is higher than 2003.Populational influence factors also have heterogeneity among provinces, and that population coefficients in 2006 is higher than 2003. (3) For all of the influence factors, GDP is a significant factor to affect carbon emissions. The evident regions affected by GDP are transferred from western to central and eastern regions in 2003 while those evident regions are transferred back to western regions in 2010. This variation has convincingly proven the complicated relations between carbon emissions and economic growth. To achieve carbon emission reduction effectively, it is significant to adjust economic structure development and improve the energy utilization efficiency. © 2013 Asian Network for Scientific Information.