基于非期望输出DEA模型的各省用水效率评价  

Water Use Efficiency Assessment of Chinese Provinces Based on a Data Envelopment Analysis Model with Undesirable Outputs

作　　者：张旭腾 于静洁[2,3] 

机构地区：[1]清华大学,北京 [2]中国科学院大学,北京 [3]中国科学院地理科学与资源研究所陆地水循环及地表过程重点实验室,北京

出　　处：《水资源研究》2017年第2期138-147,共10页Journal of Water Resources Research

基　　金：中国科学院大学生创新实践训练计划项目;国家重点研发计划项目(2016YFC0401402)

摘　　要：用水效率是水资源利用的核心,科学评价用水效率对提高区域可持续水资源管理具有重要的指导意义。本文应用数据包络分析（DEA）理论,基于用水的期望输出（直接经济价值）和非期望输出（污水排放污染环境）双重视角选择评价指标,构建了具有非期望输出的超效率DEA模型,对用水效率进行更为合理的评价。利用该模型,对1998～2012年全国各省逐年的用水效率进行计算,结果显示,北京、天津等13个省份用水效率水平较高;广东、广西等5个省份用水效率水平较低。分析表明,多数省份用水效率值及其相对高低受污水排放的影响,减少排放可提高效率。此外,通过投影分析,确定当前与效率最优状态下各指标之间的差距,为各指标的合理调整提供依据。本文研究结果通过文献调研得以印证,可对用水效率的评价和提高提供合理指导。Water use efficiency is the core issue of sustainable water resources management, which needs to be properly assessment. This paper uses the data envelopment analysis （DEA） theory and builds a super DEA model with undesirable outputs by choosing indicators including both the desirable outputs （the direct economic value） and the undesirable outputs （the sewage and pollution）. Using this model, we calculated the water use efficiency at provincial level in China during 1998-2012, and calculated the difference between the actual indicators and the indicators when the efficiency is the best with the method of projection analysis. The results indicate that the water use efficiency is relatively high in 13 provinces, such as Beijing and Tianjin; and the efficiency is relatively low in 5 provinces such as Guangdong and Guangxi. The efficiency in most of the provinces is affected by the sewage which will be raised if the sewage amount decreases. In addition, projection analysis result provides the basis for efficiency improve- ment by means of indicators adjustment.

关 键 词：数据包络分析 非期望输出超效率DEA模型 用水效率 投影分析 可持续水资源管理 

分 类 号：TV213[水利工程—水文学及水资源]