排序和广义线性模型与广义可加模型在植物种与环境关系研究中的应用

排序和广义线性模型(Generalized Linear Model，GLM)与广义可加模型(Goneralized Additive Model，GAM)是研究植物种与环境间关系的重要方法。基于线性模型的排序方法应限定于环境梯度较短的植被数据。而基于单峰模型的排序方法更适用于梯度较长的情况。PCA、CA／RA系列和CCA系列是常用的排序方法。同时进行环境数据和植被数据分析的CCA系列，能清楚地得出植物种与环境间的关系。CCA改进后的DCCA和PCCA，是现今较理想的排序方法。GLM和GAM实质上是用环境变量的高阶多项式来拟合植物种与环境变量的关系。GLM和GAM扩展了植物种与环境变量之间的关系模型，能深入地探讨植物种与环境间的关系。GLM主要是模型决定的，而GAM主要取决于原始数据。一般来说，排序能得出研究区域的主要环境梯度，提供了物种聚集和植物群落的概略描述。GLM与GAM对于深入研究单个植物种与环境间的关系具有优势。在实际研究中，两种方法结合使用能互补不足。

Application of ordination and GLM/GAM in the research of the relationship between plant species and environment

Ordination and GLM/GAM are both important methods to explore the relationship between plant species and environment.Generally,ordination methods,based on linear model,are suitable for dealing with the vegetation within a comparatively short environmental gradient,while those based on unimodal model are more suitable for tackling with the vegetation distributing along a longer gradient.PCA,CA/RA series and CCA series are the most commonly used ordination methods.Involving the environmental data in the analysis,CCA series could demonstrate the relationship between plant species and environment clearly.After modification,DCCA and PCCA are considered to be ideal ordination methods at present.Essentially,General Linear Model (GLM) and General Additive Modal (GAM) were using the high power polynomials of environmental variables to fit the response of plant species to environment.So,the models for delineating the relationship between species and environment have been enhanced through the introduction of GLM and GAM in recent years.GLM is basically driven by model,while GAM is more driven by data.In comparison,ordination can obtain the main environmental gradient and provide the general information of species assembling and plant community in the study area.GLM and GAM have advantages in probing deeply into the relationship between plant species and environment.In practice,the two kinds of methods can yield better results when used in combination.