A multivariate statistical approach to spatial representation of groundwater contamination using hydrochemistry and microbial community profiles.

Managers of landfill sites are faced with enormous challenges when attempting to detect and delineate leachate plumes with a limited number of monitoring wells, assess spatial and temporal trends for hundreds of contaminants, and design long-term monitoring (LTM) strategies. Subsurface microbial ecology is a unique source of data that has been historically underutilized in LTM groundwater designs. This paper provides a methodology for utilizing qualitative and quantitative information (specifically, multiple water quality measurements and genome-based data) from a landfill leachate contaminated aquifer in Banisveld, The Netherlands, to improve the estimation of parameters of concern. We used a principal component analysis (PCA) to reduce nonindependent hydrochemistry data, Bacteria and Archaea community profiles from 16S rDNA denaturing gradient gel electrophoresis (DGGE), into six statistically independent variables, representing the majority of the original dataset variances. The PCA scores grouped samples based on the degree or class of contamination and were similar over considerable horizontal distances. Incorporation of the principal component scores with traditional subsurface information using cokriging improved the understanding of the contaminated area by reducing error variances and increasing detection efficiency. Combining these multiple types of data (e.g., genome-based information, hydrochemistry, borings) may be extremely useful at landfill or other LTM sites for designing cost-effective strategies to detect and monitor contaminants.