植物生态学报 ›› 2011, Vol. 35 ›› Issue (12): 1256-1270.DOI: 10.3724/SP.J.1258.2011.01256

• 研究论文 • 上一篇    下一篇

山西霍山油松林的物种多度分布格局

高利霞, 毕润成, 闫明*()   

  1. 山西师范大学生命科学学院, 山西临汾 041004
  • 收稿日期:2011-05-09 接受日期:2011-09-16 出版日期:2011-05-09 发布日期:2011-12-15
  • 通讯作者: 闫明
  • 作者简介:*(E-mail:mycorrhiza@sina.com)

Species abundance distribution patterns of Pinus tabulaeformis forest in Huoshan Mountain of Shanxi Province, China

GAO Li-Xia, BI Run-Cheng, YAN Ming*()   

  1. College of Life Science, Shanxi Normal University, Linfen, Shanxi 041004, China
  • Received:2011-05-09 Accepted:2011-09-16 Online:2011-05-09 Published:2011-12-15
  • Contact: YAN Ming

摘要:

物种多度格局分析对理解群落结构具有重要的意义。该文首次选用描述种-多度关系的生态位模型(生态位优先模型NPM、分割线段模型BSM、生态位重叠模型ONM)、生物统计模型(对数级数分布模型LSD、对数正态分布模型LN)以及中性理论模型NT, 对山西霍山油松(Pinus tabulaeformis)林的物种数量关系进行了拟合研究, 并采用卡方(χ2)检验、Likelihood-ratios (L-R)检验、Kolmogorov-Smirnov (K-S)检验和赤池信息量准则(AIC)选择最适合模型, 结果表明: (1)描述乔木层物种多度格局的最优生态位模型为NPM (3种检验方法均接受该模型, p > 0.05, 且该模型具有最小的 AIC值), ONM的拟合效果次之, 不服从BSM; 三种生态位模型均可较好地拟合灌木层物种多度格局; ONM是草本层最佳生态位模型, BSM、NPM拟合效果较差; LSD可以描述油松林各层物种多度结构; LN可以很好地解释灌草层物种数量关系; NT不能解释油松林任何层次的物种多度结构。(2)霍山油松林乔木层和灌木层的物种丰富度和物种多样性均明显小于草本层; 该群落物种富集种少而稀疏种多, 且群落的均匀度相对较小。(3)从该区油松林种-多度分布来看, 同一个模型可以拟合不同的物种多度数据, 相同的数据可以由不同的模型来解释。因此, 研究森林群落物种分布时, 应采用多个模型进行拟合, 同时选用多种方法筛选最优模型。

关键词: 中性理论模型, 生态位模型, 油松林, 种-多度曲线, 物种多度分布, 种-多度关系, 统计模型

Abstract:

Aims Determination of species abundance distribution is important in research on species diversity. Our major objective was to determine species abundance distribution models to advance understanding of distribution mechanisms and to assist preservation of biological diversity.
Methods Based on data collected from field surveys, we examined the species abundance patterns of tree, shrub and herb layers in Pinus tabulaeformis forest in Huoshan Mountain, Shanxi Province. We used niche preemption (NPM), broken stick (BSM) and overlapping niche (ONM) models and two species abundance statistical models (log-series distribution model (LSD) and log-normal distribution model (LN)) and neutral theory model (NT). The simulation effects were verified by Chi-square tests, Likelihood-ratios (L-R) tests, Kolmogorov-Smirnov (K-S) tests and Akaike Information Criterion (AIC).
Important findings The best niche model for the tree layer is NPM, because it had the smallest AIC value and no significant difference (p > 0.05) between the theory predictions and observed species abundance distributions. The next best is ONM; BSM does not fit the tree layer. All three niche models are suitable for the shrub layer. ONM is the best for the herb layer, followed by BSM and NPM. LSD is good for understanding distribution mechanisms in Pinus tabulaeformisforest. LN can fit both the herb and shrub layers, but not the tree layer. NT cannot explain any layer. These findings indicate that the community has relatively few dense species and more sparse species, the species richness and diversity indices of the tree and shrub layers are much smaller than those of the herb layer, and the community evenness is relatively smaller. We conclude that one model cannot fit different data but that more than one model can fit the same data even for the same layer. Therefore, we should choose different models to study the species abundances of forest communities. We suggest that these methods might be useful for the protection of the biodiversity of forest dominated by Pinus tabulaeformis.

Key words: neutral theory model, niche model, Pinus tabulaeformis forest, species-abundance curve, species-abundance pattern, species-abundance relation, statistical model