基于DEA-ESDA的农业生态效率评价及时空分异研究

采用考虑非期望产出的SBM(Slack-Based Measure)模型,结合探索性空间数据分析方法（ESDA）,对甘肃省各县（区）2000~2014年的农业生态效率及空间分布格局进行实证分析。结果表明：① 考虑非期望产出和不考虑非期望产出的农业生态效率之间存在着明显差异,而考虑非期望产出的农业生态效率结果更符合实际情况。② 在研究期内甘肃省农业生态效率呈持续上升态势,但整体效率偏低;空间变化趋势呈现自西向东、自南向北递减态势,但南北差异小于东西差异。③ 从全局空间分异来看,除2001~2003年外,甘肃省各县（区）农业生态效率呈现正相关特征,空间集聚现象显著;从局部空间分异来看,处于空间正相关地区数量在研究期内逐渐下降,区域农业生态效率的集聚现象减弱。农业生态效率高高（HH）集聚主要分布在河西走廊及甘南地区,空间上具有明显的向邻接地区扩散的趋势;低低（LL）集聚主要分布在甘肃省中部地区,空间分布较为稳定。

Evaluation of Agricultural Ecological Efficiency and Its Spatial-temporal Differentiation Based on DEA-ESDA

The evaluation of agricultural ecological efficiency will help to understand the regional agricultural economy development status. In this paper, the agricultural water footprint and agricultural grey water footprint were respectively introduced into input indexes and undesirable output indexes. Then the SBM model of considering the undesirable outputs of DEA was used to calculate the agricultural ecological efficiency, and the Exploratory Spatial Data Analysis (ESDA) was applied to analyze its spatial distribution, agglomeration and changing trend from 2000 to 2014 in all counties (districts) in Gansu Province. The results showed that: 1) The agricultural ecological efficiency values of considering undesirable outputs were lower than the values of no consideration of undesirable outputs, and the farmer were more correspond with practical situations. 2) During 2000-2014, the agricultural ecological efficiency values were low on the whole and showed an ascending tendency in Gansu. And in spatial change, the values showed a decreasing tendency from west to east and from south to north, but the difference of values from south to north was less than that of values from west to east. 3) From the global spatial difference of efficiency values, except for 2001, 2002, and 2003, the agricultural ecological efficiency values showed the positive correlation and have remarkably agglomerated in space distribution in all counties (districts) in Gansu. From the local spatial difference, the quantity of counties with positive correlation was gradually decreasing and the spatial agglomeration of the regional agricultural ecological efficiency was reduced during the period of study. The regions with High-High (HH) correlation pattern mainly located in the Hexi Corridor and Gannan areas, and they have an obvious diffusion trend toward the adjacent areas on the space. The regions with Low-Low (LL) correlation pattern mainly located in the middle area of Gansu and the spatial distribution is relatively stable. The above evaluation results could reflect well the development condition of agricultural economy of Gansu Province in the last decade, and they could provide an important guiding and reference for realizing sustainable agricultural development in Gansu Province.