Abstract: It is important to investigate the characteristics and mechanisms of energy-related carbon emission and provide the meaningful reference for formulating the rational strategy of carbon emission reduction. The aim of decomposition analysis is to reveal the importance of distinct components or factors related to historical carbon emission data. A calculation method was used to compute the values of energy-related carbon emission and analyze the historical change of carbon emission of 6 sectors in Gansu Province during the period of 1995-2009. It was found that the total energy-related carbon emission was increased from 9.46×10⁶ t in 1995 to 1.57×10⁷ t in 2009, the average annual increase rate was 3.66%, and it was lower than the average in China and also the values of some provinces and big cities in east China, such as Beijing, Shanghai and Jiangsu. Extended Kaya identity was used to develop a factor decomposition mode and decompose the change of total energyrelated carbon emission into the components of economic scale, population size, industrial structure, energy intensity and energy structure. Based on the time series decomposition of Logarithmic Mean Divisia Index (LMDI), the change of energy-related carbon emission during the period of 1995-2009 in Gansu Province was analyzed. Empirical results demonstrated that economic scale was dominant in the increase of carbon emission in Gansu Province, then the population sizes and industrial structure. Sectoral decomposition showed that the effect of economic growth on carbon emission increase was more significant than that of industry and traffic sectors. It was found that the main factors of restricting the increase of carbon emission are the decrease of energy intensity and the change of energy structure. Some policy suggestions were put forward for reducing carbon emission in Gansu Province in the future．

Key words: carbon emission, energy consumption, reduction of carbon emission, economic scale, Gansu Province