2.5.3 Distributional Heterogeneity
Gy's distributional heterogeneity (DH) is a measure of particle distributions that can take the form of grouping or segregation. Particles may segregate, that is, separate into layers. Segregation is often a result of gravity. The most common example is jiggling a jar of dry soil, causing finer particles to migrate toward the bottom, while the larger particles end up at the top of the soil mass. Figure 2-11 illustrates the two types of DH that Gy described.
These distributional heterogeneities cause grouping and segregation error (GSE). GSE can occur at all spatial scales: within a sample (e.g., within a jar of soil) or within a field population. Note that a jar of soil is both a sample and a population. It is a sample of the field population, but it itself becomes a population when it arrives in the laboratory. That jar is the population from which a representative analytical subsample needs to be taken. If segregation has occurred (e.g., fines at the bottom and the coarser particles at the top of the soil sample jar) a sampling error is committed if the analytical subsample is taken by scooping off the top. This all-too-common sampling error easily leads to decision errors that a site is "clean" when it actually may not be.Grouping and segregation error is controlled by collecting a sufficient number of increments.
It might appear that just mixing the sample solves the problem. Unfortunately, for soil samples, common forms of mixing such as cone and quartering methods can be ineffective and may actually increase GSE (Gerlach and Nocerino 2003). Likewise, attempting to "mix" the parent matrix, such as with a backhoe, is ineffective. A good way to reduce the effects of DH is an incremental sampling approach, where enough increments are collected so that the resulting recombined large-volume sample contains the particle ratios present in the volume of matrix that was sampled.
and segregation (B) of particles.
Source: USEPA 2002e.