Abstract:

With the hypothesis that carbon emission is proportional to energy consumption,this research calculated gross and per capita carbon emissions of 30 provinces in China from 1987 to 2010,and divided provincial carbon emissions into five types by using Kmeans clustering analysis.When provincial data curves of carbon emissions over past 20 years were drawn and analyzed,it is found that there was a dramatic change in the developing trend of carbon emissions around 2002.This phenomenon implicates that the data curve of carbon emissions before 2002 could not characterize the future trends.Therefore,data of carbon emissions from 2002 to 2010 were employed,based on the Logistic model,to forecast and analyze provincial carbon emissions from 2011 to 2020.Compared forecasted data with actual ones,the margin of error in all provinces surveyed was below 7%, except for Ningxia 1458%.The average forecast error of provincial carbon emissions except for Ningxia was 622%.This result implies the accuracy of the Logistic prediction model.