基于红外相机数据的贵州高原山地环境野猪生境选择研究

Habitat selection of wild boar (Sus scrofa) in mountain area on Guizhou Plateau, China based on camera-trapping dataset

近年来野猪(Sus scrofa)在我国南方山地森林生态系统中种群数量激增、生态影响强烈,是人兽冲突的典型代表,然而对其生境选择规律仍缺乏深入研究。利用2015年7月至2020年1月的长时红外相机监测数据,对贵州高原几处环境中野猪的生境选择进行了研究,共得到野猪利用样方201个,非利用样方121个。(1)Vanderploeg和Scavia选择指数分析表明,野猪喜爱活动于坡度≤20°、乔木郁闭度>0.8和草本盖度为0.2—0.4的生境类型;不喜爱的生境类型为海拔>1600 m,坡度>25°,草本盖度<0.1,距道路距离≤100 m。(2)野猪生境资源选择函数为Logit(P)=3.226-海拔×0.002-坡度×0.161+乔木郁闭度×2.078+灌木平均高×0.401+草本盖度×3.566+距道路距离×0.001+距居民点距离×0.0003,选择概率为P=e^(logit(p))/(1+e^(logit(p))),受试者工作特征曲线(ROC)评估模型的预测精度为87.8%,能够较好预测野猪的生境选择。(3)利用Mann-Whitey U和Chi-Square Test检验不同样区2组样方之间野猪对各生境因子的选择差异表明,野猪的生境选择性还受到地形地貌、植被类型和人为干扰等因素影响。研究可为进一步深入开展华南地区野猪生态学研究,助力野猪冲突管理及科学保护提供有益参考。

In recent years, wild boar(Sus scrofa) has been regarded as a typical representative of human-animal conflicts resulted from the population increasing sharply and its consequences of ecological impacts in the mountainous forest ecosystem of southern China. However, few studies can be found on wild boar′s habitat selection in southern mountain area so far. In this study, a long-term camera-trapping dataset(from July 2015 to January 2020) obtained from several representative places in Guizhou Plateau were analyzed, and a total of 201 wild boar used plots and 121 random plots were obtained. 1) Vanderploeg and Scavia Selection Index analysis showed that the wild boar preferred the habitat where the slope ≤ 20°, tree canopy coverage >0.8 and herb coverage 0.2—0.4, while avoided the habitat where the slope >25°, herb cover<0.1 and the closest distance from road ≤100 m. 2) The Resource Selection Function model of wild boar′s habitat selection can be calculated as Logit(P) = 3.226-altitude×0.002-slope×0.161 + canopy density of trees×2.078 + shrub average height×0.401 + herb coverage×3.566 + distance from the nearest road×0.001 + distance from the nearest residential area×0.0003, accordingly, the probability of this model was expressed as P=e^(logit(p))/(1+e^(logit(p))). The prediction accuracy of the model evaluated by the Receiver Operating Characteristic(ROC) can reach to 87.8%, which indicated the wild boar habitat selection behavior can be predicted well. 3) Mann-Whitney U and Chi-Square Test were used to examine the differences in the selection of wild boar for each habitat factor between the used and random plots in different study sites, the factors including topography, vegetation types and human interference also influenced the habitat selection of wild boars. Our findings could furtherly enrich the the ecology of wild boars living in southern China, and would be also a useful reference toward the managements on human-wildlife conflicts and conservation.