---

作 者 ：骆颖,张明明,刘振生*,李志刚,胡天华,王继飞

期 刊 ：生态学报 2010年 30卷 14期 页码：3744~3751

关键词：岩羊；马鹿；同域分布；生境选择；逐步判别分析；典则判别函数；贺兰山；

Keywords:blue sheep (Pseudois nayaur), red deer (Cervus elaphus alxaicus), sympatry, habitat selection, stepwise discriminant analysis, canonical discriminant analysis, Helan Mountains,

摘 要 ：2007年和2008年9—12月，在宁夏贺兰山国家级自然保护区和内蒙古贺兰山国家级自然保护区利用痕迹检验法和直接观察法对同域分布的岩羊和马鹿发情季节生境选择进行研究。通过在选定的15条沟段里进行调查，分别测定了岩羊和马鹿的177个和154个样方的18种生态因子。结果表明，发情季节岩羊和马鹿在植被类型、地形特征、优势乔木、坡向、坡位、灌木密度、灌木高度、植被盖度、坡度、海拔高度、距水源距离、人为干扰距离、距裸岩距离和隐蔽级等生态因子选择上存在极显著差异（P < 0.01），乔木高度存在显著差异（P < 0.05），而其余生态因子无显著差异（P > 0.05），相对于马鹿选择的生境而言，岩羊发情季节更偏爱选择地势陡峭的山地疏林草原带，优势乔木以较高的灰榆为主，灌木密度低但高度较高，植被盖度较低，坡度较大的阳坡中上部，海拔较低，接近水源和裸岩，隐蔽程度低的生境。典则系数显示发情季节岩羊和马鹿的生境选择几乎完全分离，Wilk’s λ值也显示发情季节岩羊和马鹿的生境选择具有很高的差异性（Wilk’s λ = 0.123，χ2 = 679.172，df = 10，P < 0001）。发情季节岩羊和马鹿的Fisher线性判别函数分别为：F岩羊＝3.638×乔木高度＋0.242×乔木距离＋7.766×灌木高度＋0.663×灌木距离＋0.232×植被盖度＋0.191×坡度＋0.001×距水源距离＋0.008×距裸岩距离＋0.307×隐蔽级－31.078；F马鹿＝4.850×乔木高度＋0.321×乔木距离＋12.024×灌木高度＋0.929×灌木距离＋0.192×植被盖度＋0.482×坡度＋0002×距水源距离－0.001×人为干扰距离－0.003×距裸岩距离＋0.511×隐蔽级－50.787。逐步判别分析表明在区分发情季节岩羊和马鹿生境方面有一系列生态因子发挥作用，依照贡献值的大小依次为坡度、人为干扰距离、隐蔽级、距裸岩距离、乔木高度、乔木距离、距水源距离、灌木高度、灌木距离和植被盖度，由这10个变量构成的方程对发情季节岩羊和马鹿生境的正确区分率达到99.7%。

Abstract:Resource partitioning is thought to be differential use of multiple resources, such as food and space. Knowledge about habitat relationships among sympatric herbivores is important to examine resource use. Habitat selection of sympatric blue sheep (Pseudois nayaur) and red deer (Cervus elaphus alxaicus) was studied in the Helan Mountains on the border between Ningxia Moslemism and Inner Mongolia Autonomic region from September to December 2007 and 2008. We collected data by direct observation and examination of fresh sites used by blue sheep and red deer in 15 drainages distributed throughout the study area. Eighteen ecological factors from 177 sites used by blue sheep and 154 sites used by red deer were measured on 21 transects established along the ridge in the whole study area. During rutting seasons, habitat use of blue sheep and red deer differed in most ecological factors, i.e., vegetation type, landform type, dominant tree, aspect, slope location, shrub density, shrub height, coverage, slope, altitude, distance to water resource, distance to human disturbance, distance to bare rock, hiding cover, and tree height. Compared with habitat selected by red deer, blue sheep selected montane steppe dominated by Ulmus glaucescens, steeper slope, taller trees, less shrubs, taller shrubs, less vegetation cover, lower altitude, sunnier aspects, lower hiding cover, closer to water resource and bare rock. In semi-arid and rugged regions like Helan Mountains, distinct phenology promoted the utilization of different habitats by ungulates. Canonical scores indicated that habitats used by blue sheep were most separated from red deer during rutting periods. The Wilk′s lambda exhibited a highly significant difference in blue sheep and red deer habitat selection during rutting periods (Wilk′s λ = 0.123，χ2 = 679.172，df = 10，P < 0.001). The Fisher linear function discriminated blue sheep habitats is: 3.638×trees height＋0.242×distance to the nearest trees＋7.766×shrub height＋0.663×distance to the nearest shrubs＋0.232×vegetation coverage＋0.191×slope＋0.001×distance to water resource＋0.008×distance to bare rock＋0.307×hiding cover-31.078. The Fisher discriminant function of red deer habitat is: 4.850×trees height＋0.321×distance to the nearest trees＋12.024×shrubs height＋0.929×distance to the nearest shrubs＋0.192×vegetation coverage＋0.482×slope＋0.002×distance to water resource－0.001×distance to human disturbance－0.003×distance to bare rock＋0.511×hiding cover－50.787. Discriminating variables that improved a stepwise discriminant model included (in order of importance) slope, distance to human disturbance, hiding cover, distance to bare rock, trees height, distance to the nearest trees, distance to water resource, shrubs height, distance to the nearest shrubs, and vegetation coverage. Predicted accuracy of the model in classifying blue sheep and red deer habitats was 99.7%. These results suggested that differences in habitat selection were likely to have contributed to the coexistence of blue sheep with red deer during rutting periods (a period of food scarcity).

---