黄泥河自然保护区狍冬季卧息地选择

2009年12月到2010年1月,在黄泥河自然保护区采用样线法对狍冬季卧息地选择进行研究.在研究中共设置了47条样线,调查了72个狍利用样方和109个对照样方,评价了15类生态因子对狍冬季卧息地选择的影响.研究结果表明:在黄泥河自然保护区冬季,狍在卧息时喜欢选择平均海拔在591 m,位于阳坡中坡位上雪被较浅、食物丰富度、灌丛盖度、郁闭度和隐蔽水平都较高的针阔混交林生境,尤其喜欢在针阔混交林中的针叶树下卧息,避开选择阴坡、针叶林和裸岩.逻辑斯蒂回归分析结果表明:食物丰富度、针叶树、雪深、裸岩和海拔是影响黄泥河自然保护区冬季狍卧息地选择的主导因子,林型、坡向和隐蔽水平是次要因子.由这7个变量组成的回归模型为:Z=32.628+11.675×坡向(1) +9.741×坡向(2)-5.486×林型(1)-7.933×林型(2)-7.496×裸岩(1)-9.906×针叶树(1)-0.043×海拔+0.170×隐蔽水平+0.220×食物丰富度-0.429×雪深.模型选择利用概率为P(z)=ez/1+ez,整体正确预测率为96.1％.

Winter bed-site selection by roe deer (Capreolus capreolus) in Huangnihe Nature Reserve

Huangnihe Nature Reserve is located in the northwest of Dunhua City, Yanbian Korean Autonomous Prefecture, Jilin Province. As a provincial level nature reserve, it was built to protect the north temperate mountain forest ecosystems and a variety of rare and endangered wild animals and plants. According to the population and distribution of roe deer (Capreolus capreolus) in Huangnihe Nature Reserve, line transect method was used in the region which roe deer move frequently, i.e., sample points was selected randomly on the forest map. And then 47 line transects was set along with the vertical direction of the contour. Span is above 1 500 m, length is 3-5 km of the line transect which covered the major vegetation types of protected areas.Because of heavy snow in the high altitudes regions, most of line transects were below 900 m above sea level, only one transect was above 900 m, the highest point of the transect was 943 m, the snow was 86 cm thick. Roe deer lying track was set as the center for a 2 m × 2 m and a 10 m × 10 m quadrats when it was found during the field investigation, and the quadrat center was located with GPS. At the same time, a 2 m × 2 m and 10 m × 10 m quadrats was set every 1000 m of walking as control during the field investigation. If there is a roe deer lying track, then control quadrats statistics will be give up, just for lying track quadrats statistics.Seventy two quadrats used by roe deer and 109 control quadrats were measured during 47 line transect surveys with 179.2 km total length and, the effects of 15 ecological factors, namely elevation, slope direction, gradient, slope location, canopy, thickets coverage, hiding level, food abundance, distance far from human disturbance, diameter at breast height, forest type, conifer, snow depth, fallen tree and bare rock, on winter bed-site selection by roe deer were evaluated in this research totally. Comparing used quadrats with control quadrats through Baileys Confidence Interval Discriminate combining with two-sample Mann-Whitney rank test, the results indicated that significant differences were observed in selection of elevation, slope direction, slope location, canopy, thickets coverage, hiding level, food abundance, forest type, conifer, snow depth and bare rock by roe deer. Roe deer preferred the mixed broadleaf-coniferous forest habitats located in sunny middle slope, average elevation of 591m, lower snow depth, adequate food resources, thickets coverage, canopy density and hiding conditions, especially preferred conifers in mixed forest for rest while avoided habitats with shady slope, typical coniferous forest and bare rocks. Half sunny slope, broad-leaved forest, gradient, distance far from human disturbance and diameter at breast height as well as fallen tree were selected by roe deer randomly. The Binary Logistic Regression demonstrated that food abundance, conifer, snow depth, bare rock and elevation were the dominant factors leading important roles during the process of winter bed-site selection by roe deer, and forest type, slope direction and hiding level were the secondary factors. A Binary Logistic Model could be constructed by these 7 factors as follow: z= 32.628+11.675× slope direction (1)+9.741× slope direction(2) -5.486× forest type(1) -7.933× forest type(2)-7.496× bare rock(1) -9.906× conifer(1) -0.043 ×elevation+0.170×hiding level +0.220×food abundance -0.429×snow depth. The selection probability was that: P(z)= e^z/1+e^z. The model could predict correctly with 96.1% of overall.