C.1.3 Comparison of ISM and Discrete Soil Data

Tables C.1-1a, C.1-1b and C.1-1c provide a summary of the discrete sample data for PCBs for each targeted interval. A statistical evaluation of discrete vs. ISM sample data is currently under way. One objective of the review is to compare estimates of the mean concentration of PCBs in the DU soil based on a specific number of discrete samples vs. one to three multi-increment samples drawn from the same data set. Examples of the types of questions to be addressed in the evaluation include the following:

  • When is the mean estimated by the ISM samples better than the mean estimated from a randomly selected set of discrete data?
  • Is the mean estimated by a 30-point ISM sample (or set of triplicates) better than 5 discrete samples (or 10, 15, 20 samples, etc.)?

Specific information that will be determined from the data set is as follows:

  • best estimate of “true” mean PCB concentration based on all 93 discrete data points (i.e., adjusted with respect to data variability)
  • range of mean estimated by any randomly collected, 30-point ISM sample
  • range of mean estimated by any randomly collected set of 30-point, triplicate ISM samples (adjusted with respect to data variability)
  • range of mean estimated from any randomly collected set of “X” (e.g., 5, 10, 15, 20, 25, 30) number discrete samples (adjusted with respect to data variability)

For the Kure Atoll data set, the objective was to determine the equivalent number of discrete samples to a triplicate set of 30- to 50-point ISM samples. This information will help to determine the cost-effectiveness of collecting ISM samples over discrete samples.