Battelle Develops Statistical Design for Evaluating Remediation Technologies

Many companies, including Battelle, are evaluating new methods for remediating soil and groundwater. Prior to becoming accepted methods for site remediation, the abilities of these new technologies must be demonstrated. Battelle has considerable experience in designing technology demonstration studies using its expertise in technical remediation and statistical experimental design.

The use of statistics in evaluating remediation methods is extremely important because it leads to scientifically defensible results that can be presented to regulators and the general public. Remediation methods that are demonstrated to be effective when evaluated using a program based on a strong experimental design are more likely to be considered as acceptable methods for cleaning up a contaminated site.

Recently, Battelle demonstrated its statistical remediation proficiency in a field study to evaluate three different in-situ remediation technologies for the Interagency DNAPL Consortium. The study is currently being conducted at Launch Complex 34, NASA, Cape Canaveral, Fla., at a site where dense non-aqueous phase liquids (DNAPLs) have been found in soil and groundwater. Battelle and the U.S. Environmental Protection Agency’s Superfund Innovative Technology Evaluation (SITE) Program are overseeing the technical performance assessment of the three technologies:

• Permanganate oxidation
• Six-phase heating (SPH^TM)
• Steam injection/extraction.

Battelle’s design of the evaluation study is based on a primary study objective of estimating the proportion of DNAPL that is removed or destroyed. Three, 50-ft. X 75-ft. rectangular test plots were identified; one for each treatment. In order to estimate the proportion of DNAPL removed or destroyed, side-by-side pre-treatment and post-treatment samples are planned. Samples are to consist of soil cores up to 45 ft. in depth, split into two-foot sections for analysis.

Preliminary data obtained during site characterization were used to estimate variability in the DNAPL concentrations. Limits for decision errors were set, and the Data Quality Objectives (DQO) Process was applied to determine the number of samples that would be needed to meet the DQOs. The final design that was selected required 12 pairs of pre- and post-treatment cores for each test plot. Several potential methods for choosing the sample locations within each test plot were examined, and a systematic unaligned sampling procedure was selected because it is statistically efficient and provides complete coverage of the test plot.

Currently, the field demonstration of the permanganate oxidation technology has been completed, and a demonstration of the SPH^TM technology is under way. Steam injection/extraction is expected to start in the third test plot soon. Once all three demonstrations are completed, the results for the three technologies will be evaluated.

For more information about this project, contact Arun Gavaskar at (614) 424-3403 or gavaskar@battelle.org. For more information about statistical sampling designs, such as the one used at Cape Canaveral, contact Steve Naber at (614) 424-3536 or nabers@battelle.org.