阿拉善荒漠啮齿动物集合群落实证研究

当生态学家探求在破碎化的栖息地中,群落物种的共存机制、多样性、局域尺度的性质和过程被放到更广阔的时空框架内时,就出现了"集合群落"这一概念。Leibold提出了集合群落概念,他们将一个集合群落定义为局域群落集,这些群落由各个潜在的相互作用的物种的扩散连接在一起。集合群落理论描述了那些发生在集合群落尺度上的过程,并且提出思考关于物种相互作用的新方法。集合群落概念为群落生态学提供了一个新的革命性的范式,集合群落研究的最基本问题是同一系统中多物种共存的机理、多样性的形成原因与维持机制。该范式强调区域范围内群落中的综合变异,强调环境特证和栖息地之间通过扩散调节的生物相互作用和空间变化。Leibold等提出了解释集合群落结果理论上的4个生态范式,即(1)中性理论;(2)斑块动态理论;(3)物种分配理论;(4)集团效应理论。之后有大量有关检验这4种生态理论的研究,但是有关陆地脊椎动物系统的集合群落的研究较少。2010—2012年,通过在内蒙古阿拉善荒漠景观中的8个固定样地中,对啮齿动物、栖息地环境因子进行调查。利用冗余分析和偏冗余分析,评估环境特征和空间特征对物种组成的影响。结果表明,环境特征独自解释72.8%的啮齿动物物种组成变化,空间特征独自解释33.8%的物种组成变化,环境特征和空间特征共同解释86.5%的啮齿动物物种组成变化,结果显著(P=0.032);去除环境特征之后,空间特征解释13.7%的变化(P=0.246),结果不显著;去除空间特征之后,栖息地变化解释52.7%的变化(P=0.016);环境特征和空间特征的交互作用解释20.1%的物种组成的变化,该区域啮齿动物群落构成集合群落,物种共存中环境特征起着主导作用,由物种分配理论解释该集合群落结构。

An empirical research of rodent metacommunities in Alashan Desert

Introduction of the metacommunity concept has made a substantial contribution to better understanding of community composition and dynamics of a fragmentized habitat at a regional level. Leibold and Miller defined metacommunity as a set of local communities that are linked by dispersal of multiple potentially interacting species. The metacommunity concept has provided a new and revolutionary paradigm for the community ecology by emphasizing integrated variation among communities at the regional level. This variation is accentuated by biotic interactions and spatial variation in environmental characteristics and modulated by dispersal among sites. The basic problems of metacommunity research are the multiple species coexistence mechanism, the reasons behind the diversity, and mechanism maintaining this diversity within the same system. Leibold and Miller proposed four paradigms for metacommunities: the patch-dynamic view, the species-sorting view, the mass effects view, and the neutral view, each emphasizing different processes of potential importance in metacommunities. A variety of studies verified the four ecological paradigms, but only few addressed the terrestrial vertebrates system. In recent years, as a result of human activities and natural disturbances, Alashan Desert habitat in Inner Mongolia has been largely fragmentized. This habitat fragmentation is a major threat to animal communities. From 2010 to 2012, we investigate rodent communities and habitat environmental factors at eight fixed experimental sites in Inner Mongolia''s Alashan Desert. We used a set of full and partial redundancy analyses to estimate four additive partitions of variance in rodent species composition: (a) unexplained variation, (b) variation explained by spatial characteristics, (c) variation explained by local environmental conditions, and (d) variation explained by both spatial and environmental characteristics. The results of full and partial redundancy analyses showed that habitat and spatial characteristics alone accounted for 72.8% and 33.8%, respectively, of the variation among sites regarding rodent species composition. All explained variables accounted for approximately 86.5% of the variation (P = 0.032) in rodent species composition among sites. Spatial characteristics accounted for 13.7% of the variation (P = 0.246) in species composition after removing variation shared with habitat characteristics. Habitat characteristics accounted for 52.7% (P = 0.016) of the variation in species composition after removing variation shared with spatial characteristics effect. Finally, spatial structured habitat characteristics (interaction term) accounted for approximately 20.1% of the total variation in rodent species composition among sites. The environmental component to rodent species composition was greater than the spatial component. Pure environmental characteristics accounted for significant amounts of variation in rodent species composition, and pure spatial characteristics accounted for non-significant amounts of variation in rodent species composition. The results from variance decomposition suggested that the species sorting model of metacommunity structure fits better the rodent community than the mass effect or neutral models.