Kitpati S, Fujii S, Shimizu Y
Research Center for Environmental Quality Control, Graduate School of Engineering, Kyoto University, 1-2, Yumihama, Otsu, 520-0811, Japan.
Water Sci Technol. 2005;52(9):283-94.
Lake water quality systems are extremely complex and every lake is characterized by many complicated interactions between biological, chemical and physical variables. One of the most important tasks of lake mechanism analysis is to examine the relations among important water quality variables and to describe such complicated relationships quantitatively, simply and practically. Lake Biwa was chosen as a target lake, where a periodic water quality investigation was conducted every month from 1979 at 50 sampling sites. The objective of this study is to establish an application procedure of three-ways layout data and to present a vision including its practical result, expecting that this vision may also provide a direction for further research. The results showed that (1) cluster analysis following the three-ways layout ANOVA method is very effective in presenting water quality variation, (2) correlation analysis of water quality indices in each effect of parameters and their interaction between two items verifies that their relationships in parameters are different due to different mechanisms. (3) An integrated analysis procedure was derived, potentially describing those data in forms which convey their important characteristics and reveal their significant relationship. The same approach of this analysis procedure could be possibly applied for other water quality indices in other lakes.
湖泊水质系统极其复杂，每个湖泊都具有生物、化学和物理变量之间许多复杂的相互作用。湖泊机理分析的最重要任务之一是研究重要水质变量之间的关系，并以简单、实用的方式定量描述这种复杂关系。琵琶湖被选为目标湖泊，自1979年起每月在50个采样点进行定期水质调查。本研究的目的是建立三向布局数据的应用程序，并展示其实际结果，期望这一成果也能为进一步研究提供方向。结果表明：（1）采用三向布局方差分析方法进行聚类分析，对于呈现水质变化非常有效；（2）对各参数效应及其两项之间相互作用的水质指标进行相关性分析，验证了由于不同机制，各参数之间的关系存在差异；（3）得出了一种综合分析程序，有可能以传达重要特征并揭示显著关系的形式描述这些数据。该分析程序的相同方法可能适用于其他湖泊的其他水质指标。
