【摘要】 以2002年印度野牛(Bos gaurus)印度次大陆的空间分布图和开放式生态环境数据库的30个环境变量为主要信息源,采用逻辑斯蒂回归(LR)、分类与回归树模型(CART)、基于规则的遗传算法(GARP)、最大熵法(Maxent)4种生态位模型,进行印度野牛印度大陆部分潜在生境预测。结果表明:4个预测模型总体性能相差不大,其中,Maxent模型单项指标、总体平均值均最高;年平均降水量、温度季节性变化、林木覆盖率、标准归一化植被指数、海拔高度、距主要道路距离是影响印度野牛空间分布的主要环境因子;增加林木覆盖率、减少道路修建引起的生境破碎化、保证水源是印度野牛保护的主要措施。更多还原

【Abstract】 Distribution map for gaur Bos gaurus in 2002 and 30 environmental variables from open infrastructure geographical and ecological environmental datasets were gathered,and four modeling methods of Logistic Regression(LR),Classification and Regression Trees(CART),Genetic Algorithm for Rule-Set Prediction(GARP),Maximum Entropy method(Maxent)were introduced to map potential geographic distributions for gaur in the India Subcontinent.Results show that there is not distinct difference in overall performance among four ecological niche models,in which Maxent outperformed others in two aspects of single criteria and overall average criteria.Annual precipitation,temperature seasonality,tree cover,normalized difference vegetation index,altitude,distance to main road are the environmental forcing factors.Decreasing habitat fragmentation by improving tree coverage and reducing road density,assuring water supply are the main measures for Bos gaurus conservation.更多还原