R程序包“rdacca.hp”在生态学数据分析中的应用:案例与进展

Application of“rdacca.hp”R package in ecological data analysis:case and progress

定量评估不同变量对群落组成的贡献是群落生态学分析的热点问题。但在具体的分析情景中,因子间的共线性与解释率的重叠对评估不同因子重要性造成了较大困难。基于这一问题,R程序包“rdacca.hp”通过引入层次分割法(HP)的理念,在所有可能的模型子集下为各解释变量(或解释变量组)分配单独效应,为典范分析中共线性解释变量的相对重要性评估提供了新的定量指标。目前,“rdacca.hp”包已经成为群落生态学分析的重要工具。为进一步促进用户对“rdacca.hp”包的理解与运用,该文通过引入一个分析塑造甲螨(Oribatida)群落的重要环境和空间驱动因素的实例,重点展示了使用该程序包进行典范分析的一般步骤。随后对近期应用“rdacca.hp”包开展分析的相关研究进行文献计量学分析,结果表明该程序包自上线以来已被广泛用作解决生态学、环境科学及相关学科问题的基本定量框架。最后,该文对“rdacca.hp”包的未来应用和升级进行了展望。总之,该文旨在使国内学者们进一步加深对“rdacca.hp”包的认识和应用。

Quantitative estimation of the contribution of predictary variables to community composition is a hotspot in community ecology.However,multicollinearity and joint contributions among predictors make it difficult to estimate the importance of predictor in specific analysis scenarios.To address this issue,the“rdacca.hp”package provides a new quantitative indicator by introducing the concept of hierarchical partitioning(HP)to assign individual effects for individual predictors(or groups of predictors)across all possible model subsets.The package solves the problem of estimating the relative importance of predictors with multicollinearity in canonical analysis.The“rdacca.hp”package has become an important tool for community ecological analysis.To further promote users’understanding and use of the“rdacca.hp”package,we demonstrate the general steps for using this package in canonical analysis with an example analyzing the important environmental and spatial drivers that shape the oribatid mites(Oribatida)community.Subsequently,we conduct a bibliometric analysis of recent studies using“rdacca.hp”package.The results show that,since its launch,the package has been widely used as a fundamental quantitative framework in ecology,environmental science and related disciplines.Finally,we discuss the further application and extension of the“rdacca.hp”package.In conclusion,this paper aims to advocate the understanding and application of the“rdacca.hp”package for domestic researchers.