作 者 ：黄宝荣,欧阳志云,张慧智,郑华,徐卫华,王效科
期 刊 ：生态学报 2008年 28卷 1期 页码：327~337
关键词：行政区；生态环境质量；指标；指数；评价；主成分分析；聚类分析；
Keywords:administrative divisions, eco-environmental quality, indicator, index, assessment, principal component analysis, cluster analysis,
摘 要 ：从自然条件、人类胁迫、生态环境效应和社会响应4个主题出发，构建了包含28个指标的中国省级行政区生态环境可持续性评价指标体系,并通过专家调查和层次分析法为各主题层中的指标赋权。分别用各主题层中的指标和对应的权重，采用加权平均法构建自然条件（NSI）、人类胁迫（HPI）、生态环境效应（EEI）和社会响应（SRI）4个分指数，然后用这4个分指数通过灰色关联法构建综合生态环境可持续性指数（CEI），对我国各省级行政区生态环境可持续性进行评价。同时论文通过主成分分析法和相关分析法分析造成各指数地域差异的主导因素和评价中指标赋权的合理性。最后以4个分指数作为变量，采用快速样本聚类法对我国31个省级行政区进行聚类分析。评价结果表明，2003年我国各生态环境指数存在显著的地域差异。自然条件从东南到西北呈明显的由优到劣的过渡趋势，降雨量、年均气温、森林覆盖率和水网密度4个指标是导致地域差异的主导因素；人类胁迫东部发达地区明显高于西部和中部地区，同一个地区经济发展水平存在显著的正相关；生态环境效应西北地区最为显著，同各地区自然条件存在显著的负相关，在人类活动干扰下，自然条件差的地区更容易发生生态损害与退化；社会响应无明显的地域分析规律，各地区经济发展水平对其有一定影响，但并非决定性因素。区域综合生态环境可持续性由区域NSI、HPI、EEI和SRI共同决定， 综合评价得到2003年我国各省级行政区生态环境可持续性排序情况，其中生态环境可持续性最强的几个地区为海南、云南、广西、江西等地，最弱的几个地区为内蒙古、宁夏、上海、西藏、新疆等地。最后，聚类分析把我国31个省级行政区聚为生态环境特点各异的六大类，可以促进我国生态环境的分类管理。
Abstract:Agenda 21 emphasizes the role of local authorities as an important resource in pursuit of sustainability. Key aspects in this process are the collection of information about the state of the environment and the integration of this information at the various levels of local governance to highlight useful patterns for policy planning. As it is very difficult to evaluate the environmental performance on the grounds because of too many environmental indicators, i.e. too much information, we should reduce the number of indicators by aggregating them to one or several composite environmental indexes to make the environmental information more accessible.
Based on “naturally occurring state; human pressures; environmental effects; and social responses” four environmental sustainability themes, this paper developed an indicator system including ten contributing factors and twenty-eight indicators for assessing the environmental sustainability of each province in China. In this categorical system, indicator weight was determined through analytic hierarchy processes and expert survey methods. Four sub-indices——including natural state index (NSI), human pressure index (HPI), environmental effect index (EEI), and social response index (SRI)——were aggregated through additive weighting method. Then, the composite environmental sustainability index (CEI) of each province was calculated by averaging the grey relation performance correlations between the four sub-indices and their hypothetical ideal point. Higher values of CEI demonstrated that the environment of a local province is more sustainable. At the same time, indicators in each theme were analyzed by correlation and principal component analysis (PCA) in search of meaningful underlying variables which derive regional differentiation of each sub-index. Finally, the paper conducted a cluster analysis, which identifies statistically related groups of provinces based on the similarity of sub-indices, a step that would likely prove very helpful in identifying appropriate provinces against which local environmental sustainability and policies can be benchmarked and compared.
The results demonstrated that in each province there were obvious regional differentiations of natural conditions, human pressures, environmental effects, and social responses. The NSI in each province of East and South China was greater than West and North China, a result that was mainly determined by rainfall, air temperature, water-resource index and rate of forest cover in each province. The HPI in each developed province of Eastern China was greater than in poor provinces of Central and Western China, which demonstrated significant positive correlations with economic level, namely the GDP per capita in each province. The EEI in Northwest China was greater than for other regions, and it is negatively correlated with NSI; in regions with a poor natural state, little human pressure would lead to significant ecological effects. The SRI did not show any obvious special differentiation rules——that is, economic growth would promote social response, but wasn’t decisive factor; and social response also influenced by national policy and local development strategy. Eco-environmental sustainability in each province was determined by NSI, HPI, EEI and SRI. Through grey correlation analysis of the four sub-indices, this paper developed the CEI of each province. According to 2003 data, Hainan Province, Yunnan Province, Guangxi Autonomous Region, and Jiangxi Province rate at the top ranks of CEI. These four regions are endowed with good natural states, less human pressure, unapparent eco-environmental effects, and high social responses. At the bottom ranks of CEI are four Autonomous Regions——Neimengu, Ningxia, Tibet and Sinkiang——with poor natural states and serious eco-environmental effects, along with Shanghai Municipality which exhibits high human pressure and serious eco-environmental effects. The cluster analysis classed thirty-one provinces into six groupings which could be seen as having observable similarities and thus representing a useful point of departure for policy comparisons.
全 文 ：