稀有种处理对RDA排序结果影响的比较研究

为了解稀有种对RDA排序结果的影响,该研究以北京东灵山华北落叶松林调查数据为例,在RDA排序的基础上,对比分析了未处理稀有种RDA与剔除频度5%、盖度5%的稀有种后RDA排序结果的差异,并用蒙特-卡罗拟合检验分析了二者物种变量和环境变量之间的相关关系,以及用Spearman秩相关系数检验了对应排序轴的相关性。结果表明:(1)蒙特-卡罗拟合检验结果显示未处理稀有种RDA与剔除稀有种RDA各自对应的物种变量和环境变量之间均呈极显著相关关系;(2)从排序轴特征值对物种数据方差以及物种—环境关系解释量来看,剔除稀有种RDA前两排序轴与前四排序轴均有较高的物种-环境关系累积解释量;(3)剔除稀有种前后对应排序轴的Spearman秩相关分析结果表明,尽管未处理稀有种RDA和剔除稀有种RDA在第三轴和第四轴间存在一定的交叉,但二者对应的前四排序轴均呈极显著的一一对应关系(P0.001),相似性极高。总之,结合物种-环境关系的累积解释量及对应排序轴的相关性可知,在环境因子个数较少、研究尺度较小时,使用RDA排序揭示植物种、植物群落和环境因子之间相互作用的生态关系时,剔除稀有种前后RDA排序具有较高吻合性,只是对环境因子的解释趋势稍有差异。

Comparative study concerning the effects of rare species treatments on RDA ordination

On the basis of redundancy analysis(RDA), eliminating the rare species whose frequency and coverage are less than 5% in Dongling Mountain of Beijing and our objective was to explore the effects of rare species on RDA ordination and summarizing the using-actuality of two methods. In this paper, Monte Carlo permutation test was used to analyse their relationship between species and environmental variables, and the correlation of corresponding ordination axes of two methods was measured by Spearman''s rho correlation. The results were as follows:(1)The Monte Carlo permutation test results showed that their species and environmental variables had a significant correlation;(2)From the perspective of the cumulative percentage variance of species-environment relation, the RDA results of eliminating rare species had a higher amount of the interpretation of species-environment relationship both in the first two axes and the first four axes, therefore, it could better explain the relationship between species and the environment;(3)Two-dimensional RDA ordination diagram revealed the distribution of environmental variables, samples and species were basically the same; the results from the Spearman rank correlation of corresponding axes showed that the third axis of untreated rare species corresponded to the forth axis of eliminating rare species and the forth axis of untreated rare species corresponded to the third axis of eliminating rare species, whereas there were very prominent relationships between their first four axes(P<0.001), which revealed the same environmental gradient. In conclusion, the findings in this study suggested that the results of two methods were basically the same when the number of environmental factors was less, but some differences were existed on the explanation tendency for their environmental factor.