DIMENSION REDUCTION AND SOURCE IDENTIFICATION FOR MULTISPECIES GROUNDWATER CONTAMINATION

Development of principal composants analysis (PCA) applied to a snapshot of organic and inorganic chemicals sampled from observation wells within a contaminated shallow grounwaters. Description of sites and data (116 chemical species at 184 wells). Ranked and normalized eigenvalues (fractional variance)for the correlation matrix of chemicals. For each well a normalized group concentration index is defined. Clustering of the data on eigen-space with transition groups and primary groups.