地形因素对羚牛家域估算的影响分析——以固定核域法与最小凸边形法为例

家域特征对物种基础生物学的认识和提出相关管理及保护对策有重要作用.本研究利用地理信息系统相关软件,以4只(2♀、2 ♂)佩戴GPS无线电颈圈的四川羚牛定位数据使用固定核域法(FKE)和最小凸多边形法(MEP)对春季(2007年和2008年)家域特征进行了估算,显示地形因素对家域估算结果影响显著,表面家域(FKE 95％ =7.50 ±2.27,MCP =7.01±1.99)要显著大于平面家域(FKE 95％=5.94±1.54,t=3.31,df=3,P=0.045,MCP=5.47±1.52,df=3,t=3.43,P=0.041).其次,固定核域法(95％)与最小凸多边形法所获得的平面家域(t =0.718,df=3,P=0.524)和表面家域(t=0.612,df=3,P=0.584)差异不显著.个体间家域差异显著(df=3,F=7.226,P=0.001),组间两两比较显示,M1与F1 (P =0.001)、F2(p=0.031)、M2(P=0.02),F1与F2 (P =0.044)间的家域均有显著差异,而M2与F2 (P =0.221)、M2与F1(P=0.598)间差异不显著.两种家域估算方法对计算羚牛的家域重叠程度没有明显优劣之分(95％ Kernel VS 100％ MCP),但不同密度的核域可能得出不同的甚至是完全相反的结论,因此在选择核域密度时应尽量同时参照其他数据和其他估算方法.

Analysis of terrain induced impact on Takin(Budorcas taxicolor) home range estimation:an example from Fixed kernel estimator and Minimum convex polygon

Knowledge of the home range of wildlife is the basis of effective conservation and management.We studied takin(Budorcas taxicolor tibetana) home range in spring with two widely applied methods using GIS related software,where we compared differences between the home range estimates with and without terrain information.All four takin(2♀2♂) were fixed with GPS radio collars and the data were obtained in the spring of 2007 and of 2008.The results showed home range size derived with terrain information(FKE95%=7.50±2.27,MCP=7.01±1.99) was significantly larger than that derived without topographic information(FKE95%=5.94±1.54,t=3.31,df=3,P=0.045,MCP=5.47±1.52,df=3,t=3.34,P=0.041).The difference between estimates based on 95% Fixed kernel home range and Minimum convex polygon home range were not significant regardless of including terrain information(t=0.612,df=3,P=0.584) or not(t=0.718,df=3,P=0.524).However,we also found significant differences of home range size between individuals(One way ANOVA,df=3,F=7.226,P=0.001).The post hoc test between individuals found significant differences between M1 and F1(P=0.001),M1 and F2(P=0.031),M1 and M2(P=0.02),F1 and F2(P=0.044),whereas no significant difference was detected between M2 and F2(P=0.221),M2 and F1(P=0.598).We also tested the difference of the two methods on the results of spatial overlap between individuals,but no significant difference was detected.Researchers should select appropriate smooth parameters(h)/bandwidth,because different bandwidth may derive different conclusions of the degree of home range overlap,or even contradictory results.