van den Brink Paul J, den Besten Piet J, bij de Vaate Abraham, ter Braak Cajo J F
Alterra, Wageningen University and Research Centre, P.O. Box 47, 6700 AA, Wageningen, The Netherlands.
Environ Monit Assess. 2009 May;152(1-4):271-81. doi: 10.1007/s10661-008-0314-6. Epub 2008 May 17.
Although chemical and biological monitoring is often used to evaluate the quality of surface waters for regulatory purposes and/or to evaluate environmental status and trends, the resulting biological and chemical data sets are large and difficult to evaluate. Multivariate techniques have long been used to analyse complex data sets. This paper discusses the methods currently in use and introduces the principal response curves method, which overcomes the problem of cluttered graphical results representation that is a great drawback of most conventional methods. To illustrate this, two example data sets are analysed using two ordination techniques, principal component analysis and principal response curves. Whereas PCA results in a difficult-to-interpret diagram, principal response curves related methods are able to show changes in community composition in a diagram that is easy to read. The principal response curves method is used to show trends over time with an internal reference (overall mean or reference year) or external reference (e.g. preferred water quality or reference site). Advantages and disadvantages of both methods are discussed and illustrated.
尽管化学和生物监测常常被用于出于监管目的评估地表水质量和/或评估环境状况及趋势，但由此产生的生物和化学数据集规模庞大且难以评估。长期以来，多变量技术一直被用于分析复杂的数据集。本文讨论了目前使用的方法，并介绍了主响应曲线法，该方法克服了图形结果表示混乱的问题，而这是大多数传统方法的一个重大缺陷。为了说明这一点，使用主成分分析和主响应曲线这两种排序技术对两个示例数据集进行了分析。主成分分析得出的图表难以解读，而主响应曲线相关方法能够在一张易于阅读的图表中显示群落组成的变化。主响应曲线法用于通过内部参考（总体均值或参考年份）或外部参考（例如，理想水质或参考站点）来展示随时间的趋势。讨论并举例说明了这两种方法的优缺点。
