Contamination Distribution and Mass Estimate via Kriging – Pitfalls and Lessons

Determining the volume and mass of soil/dissolved contaminants is an important component in characterization and remediation design. The distribution and magnitude of detected concentrations are the basis from which the estimate of volume and mass are determined. Computer technologies have enhanced the capabilities of conducting the analysis by coupling three dimensional interpolation with visual imagery. Three dimensional kriging is a particularly useful interpolation technique as it not only provides a quantitative estimate but also enables an assessment of uncertainty. Evaluating uncertainty in the volume and mass estimate is an important component of the analysis as it allows quantitative assessment of the reliability. This understanding can then be incorporated into other analyses such as fate and transport modeling, delineation of treatment zone, and sizing of the remedial system. Although computer technologies have facilitated conducting these volume and mass analyses, caution should be exercised as vastly erroneous estimates can result from unforeseen conceptual and technical errors. The errors normally occur during domain selection, nondetects quantification, use of control points, and confidence interval estimate. This paper documents the application of three dimensional kriging coupled with state-of-theart visualization for a complex dissolved MTBE plume. The various pitfalls and lessons of the analyses are presented, with emphasis on how potential errors can significantly alter the results.