基于逻辑斯蒂回归模型的鹭科水鸟栖息地适宜性评价  

Assessment of ardeidae waterfowl habitat suitability based on a binary logistic regression model

作　　者：邹丽丽[1] 陈晓翔[1] 何莹[1] 黎夏[1] 何执兼[1] 

机构地区：[1]中山大学地理科学与规划学院,广东省城市化与地理环境空间模拟重点实验室,广州510275

出　　处：《生态学报》2012年第12期3722-3728,共7页Acta Ecologica Sinica

基　　金：广东省联合基金项目(U0833005)(遥感调查专题)

摘　　要：近年来湿地生态系统遭到不同程度破坏,湿地水鸟及其生存空间日益受到威胁。以香港米埔-后海湾湿地为例,收集2003年1月份与鹭科水鸟密切相关的15个自变量和鹭科水鸟实测数据作为因变量构建逻辑斯蒂回归模型,通过筛选获取9个变量因子,分别为土地利用,NDVI,坡度,降雨,TM4纹理,TM3纹理,道路密度,道路距离,人居密度。经Nagelkerke R2检验模型精度达到0.743,拟合度较高。利用模型结果快速聚类,对栖息地进行适宜性分级,分级结果与同期鹭科水鸟实测数据做拟合,精度达到77.4%。最后采集2009年1月份各变量因子数据对回归方程进行时间尺度检验,与同期实测鹭科水鸟数据拟合精度同样达到75.8%,模型具有较好的通用性。Ardeidae and their habitat are being threatened by the destruction of wetland ecosystems.Because the availability of suitable Ardeidae habitats are gradually decreasing,habitat suitability evaluations using remote sensing and measurement technology are becoming research hotspots.In this study,Mai Po and the Deep Bay Area in Hong Kong was chosen as a case study area.Based on descriptive waterfowl statistics from field observation data,we used Moran′s index to assess spatial autocorrelation.The results of this analysis indicated that the field data was randomly distributed within the study area,and that there were no duplicate records.Therefore the data was suitable for modeling.We used a logistic regression model to determine the factors associated with the waterfowls′ habitat.First,according to Ardeidae habitat dependencies,we collected data of 15 factors relevant to the waterfowl in January 2003 using Geographic Information System（GIS） and Remote Sensing technology.We randomly selected 1000 points from each factor′s grid map and used the leverage value and cook distance method to look up unusual points.Few unusual points were identified,and most of the selected points were able to be used to establish the binary logistic regression model.Nine factors were identified by the model as having an important influence on the presence/absence of Ardeidae.These factors were land use,Normalized Difference Vegetation Index,slope,rainfall,TM 4 texture,TM 3 texture,road density,road distance,and human habitation density.The Nagelkerke R2 verified that the model was a good fit（coefficient =0.774）.Aspect,digital elevation,human habitation distance,TM 2 texture,relative humidity,and temperature were simultaneously filtered out of the model,suggesting that the waterfowls′ habitat preferences were not influenced by these factors.We then used a quick clustering method to grade habitat suitability,and found that the case area could be divided into five levels.From a two figure stack in GIS,we selected 500 points fro

关 键 词：逻辑斯蒂回归模型 鹭科水鸟 栖息地 适宜性 

分 类 号：X821[环境科学与工程—环境工程]