目的 在青海省喜马拉雅旱獭监测数据的基础上,利用地理信息系统(GIS)和基于遗传算法的规则集预测(GARP)模型进行旱獭的空间分布预测.方法 整理青海省喜马拉雅旱獭的常规监测和野外全球定位系统(GPS)调查数据,利用GIS软件进行位置点的空间化处理;提取与分析包括地形、坡度、气温、降水、植被、土地利用等与喜马拉雅旱獭相关的生态环境变量;利用GARP模型和ArcGIS构建旱獭的空间分布预测模型,并进行空间分布制图与分析.结果 结合空间位置恢复和GPS点信息,共处理得到喜马拉雅旱獭位置点198处;通过GARP建模得到模型的平均遗漏误差为1.998,而优选的100个模型均具有统计学意义(x2值均> 163.03,P均<0.01);利用ArcGIS将预测的空间分布概率分为<40%、40%~<80%和80%~100%3个等级,其中模型预测概率为80%~ 100%的区域是喜马拉雅旱獭最适宜的分布地区.结论 利用GIS和GARP生态位模型可以进行鼠疫宿主动物喜马拉雅旱獭的空间分布预测,其结果比按行政区划进行的空间分布统计更加准确,可以为鼠疫防治提供重要参考.
Objective To predict the spatial distribution of Himalayan marmot using geographic information system (GIS) and Genetic Algorithm for Rule-set Production (GARP) model based on the monitoring data of Himalayan marmot in Qinghai Province.Methods Based on the data of routine monitoring and field survey by Global Position System (GPS) of Himalayan marmot,the position data was processed by spatial mapping using ArcGIS software.Ecological environment variables related to Himalayan marmot including terrain,slope,temperature,precipitation,vegetation,land use and other related variables were extracted and analyzed.The prediction model of Himalayan marmot distribution was constructed based on GARP model and ArcGIS software,and the spatial distribution mapping and analysis were carried out.Results Totally 198 points of Himalayan marmot were obtained by combining the recovery position with GPS information.The average model error of omission was 1.998 through the GARP modeling,while the optimal 100 model were highly statistically significant (all x2 〉 163.03,all P 〈 0.01).The spatial distribution of predicted probability was divided into three grades including less than 〈 40%,40%-〈 80% and 80%-100% using ArcGIS,and the area with the prediction probability of 80%-100% was the most suitable distribution area of Himalayan marmot.Conclusions The spatial distribution of plague host animal is predicted successfully using GIS and GARP ecological niche model.The result is more accurate compared to the statistic area by administrative region,which can provide important reference for plague prevention and control.