基于偏最小二乘模型的碳排放区域差异及影响因素的实证分析  

Regional Difference and Influence Factors of Carbon Emissions:Empirical Analysis Based on Partial Least Squares Method

作　　者：张先锋[1] 韩雪[1] 张庆彩[1] 

机构地区：[1]合肥工业大学,合肥230009

出　　处：《工业技术经济》2013年第7期100-109,共10页Journal of Industrial Technological Economics

基　　金：2010年度教育部人文社会科学研究规划基金项目(项目编号:10YJA790252);2012年度国家软科学研究计划项目(项目编号:2012GXS4D085);中央高校基本科研业务费专项资金项目(项目编号:2012HGXJ0080)

摘　　要：本文对我国30个省份2000~2010年主要能源的二氧化碳排放量进行测算,通过K-均值聚类分析将各省域按照"排放———效率"类型分为四大类,并运用偏最小二乘建模分析了能源消费结构、经济发展水平、产业结构、人口规模、城市化率以及碳排放强度对4类区域碳排放的影响。结果表明:4类区域的二氧化碳排放存在明显差异,人口规模、产业结构和碳排放强度是4类地区碳排放共同的驱动因素;对于高排———低效地区和低排———低效地区,经济发展水平与碳排放呈倒U型关系,高排———高效地区与低排———高效地区的经济发展水平与碳排放之间并无显著的倒U型关系;另外,能源消费结构与4类地区碳排放的相关性存在明显差异。本文对上述差异做出分析与解释,并提出了相应的政策建议。This article uses ORNL＇s method to calculate the carbon emissions in 30 pmvinees from 2000 to 2010, divide 30 provinces into four main categories in accordance with the type of ＂emission - et^ciency＂ through K - means cluster analysis, then PLS model was adopted to study the drivers of carbon emissions including energy consumption structure, the level of economic development, in- dustrial structure, population size, urbanization rote and carbon emissions intensity. The results show： obvious differences are fond among the four types of regions. Population size, carbon intensity and carbon intensity are the most important drivers of carbon emissions for all the four types of regions; inverted U- shaped can be used to discfibe the relationship between carbon emission and the economic growth in ＂HE- LE＂ and ＂LE- LE＂ areas but not suitable for ＂HE- HE＂ and ＂LE- I-I~＂ areas. There exists significant differences in the cor- relatious between the structure of energy consumption and carbon emissions of the four regions, this paper analyses the mason of the differ- ences and provides corresponding policy suggestions.

关 键 词：碳排放 聚类分析 偏最小二乘回归 区域差异 

分 类 号：F129.9[经济管理—世界经济]