When discrete sampling may be successful

The problems with making decisions about large bulk volumes of soil using discrete sample data have been discussed throughout this section. The particulate nature of soil and its interaction with contaminants, as well as the sheer volume disparity between the amount of soil analyzed and on which decisions are made, means that heterogeneity is the primary factor affecting the sampling error and thereby affecting the quality of environmental decisions.

One is most likely to make correct environmental decisions using discrete sampling in the following circumstances:

  • Low-density discrete sampling may be sufficient when the impacts of heterogeneity and sampling error on the decision are expected to be low:
    • previously collected discrete sampling data indicate that the mean (or range) of soil concentrations is well below the action level,
    • previously collected discrete sampling data demonstrate that heterogeneity is very low, or
    • the sampling goal is to obtain qualitative data, for example, when conducting in situ X-ray fluorescence (XRF) soil screening to gain initial estimates of the nature and extent of metal contamination.
  • High-density discrete sampling (roughly equivalent to the number of increments collected with ISM) can be useful when sample locations and sampling and subsampling techniques are appropriate for obtaining an unbiased estimate of the mean.
  • The volume of soil represented by the discrete sample or samples can be adequately identified. Note that that volume of soil to which discrete samples apply is often determined after the samples are collected and the data apportioned in a variety of different ways, as further discussed in Hyperlink 16.

There are other situations where discrete sampling may be preferred, even though the above conditions are not met, for instance, when (a) discrete sampling is required by regulation, (b) sample collection and/or processing may change the concentration of the sample (e.g., reactive chemicals are investigated), or (c) ISM is cost-prohibitive.