سال انتشار: ۱۳۹۰
محل انتشار: پنجمین همایش تخصصی مهندسی محیط زیست
تعداد صفحات: ۸
M Rezaei – Graduate student, Department of Civil and Environmental Engineering, Shiraz University, Shiraz, Iran
M. J Abedini – Associate Prof., Department of Civil and Environmental Engineering, Shiraz University, Shiraz, Iran.)
Recent advances in multi-sensor technology and electronics created a situation where huge amount of data (in both space and time) can be obtained in various disciplines including environmental engineering. These data are often multi-dimensional with numerous hidden properties. The univariate statistical analysis, generally implemented to treat such data, could not uncover the complex and hidden properties of them. Multivariate techniques are the key puzzle in this struggle. They are unbiased techniques which can help indicate natural associations between samples and/or variables, thus highlighting information not available at first glance. The multivariate statistical techniques and the associated exploratory spatial data analysis (ESDA) are the appropriate tools for a meaningful data reduction and interpretation of multi-constituent chemical, physical, and biological measurements. In this paper, the usefulness of such tools as principal component analysis (PCA), factor analysis (FA) and cluster analysis (CA) will be highlighted regarding their potential to discover the hidden properties of data in various branches of environmental engineering including surface water, groundwater, soil, and air. Careful review of literature showed that PCA has been widely used in geochemical applications to identify pollution sources and to separate natural versus anthropogenic activities in aqueous, soil and air environment, as well as examination of spatial and temporal patterns of contamination and also identification of the most informative monitoring stations. The associated procedure of factor analysis with its robust rotation ability could facilitate the clean up of principal components. CA is an efficient means to recognize groups of samples that have similar chemical and physical characteristics. It is concluded that multivariate statistical techniques could be considered as an efficient tool to display complex relationships among many objects and assist the decision makers in managing and controlling pollution and provide an effective overview of hot spots where intensified monitoring activities are required.