Statistics - Field Sampling: Principles and Practices in Environmental Analysis - page 147

self-evident. Air, water, and soil appear to be continuous media, however, and not related
to populations at all. In this case each sample is considered a member of a population.
One way of looking at this is to ask if the samples are part of a group of samples that are
the same or if they are from a different group of samples.
Soil scientists do define a soil individual; that is, a volume of soil that has well- and
precisely defined characteristics. Such a soil, which is 1 to 10 m2 and 1.5 to 2 meters
deep, is called a pedon, and a group of pedons in a field would be called a polypedon. It
is rare that a field to be sampled would consist of only one type of soil or pedon,
however. When soils are mapped, as in a soil survey (see Figure 1.2 in Chapter 1), the
different soils noted are not technically pedons, but soils of similar characteristics called
mapping units. Soil sampling thus is not sampling individuals in the soil science sense.
Both the atmosphere and hydrosphere are also seen to be continuous and homogenous.
When they are sampled, however, each sample is considered to be a member of a
population of samples, just as soil samples are.
Often three soil cores will be crushed and mixed to provide a sample for analysis. This
sample is also considered to be a member of a population of soil samples. This is true as
long as it is analyzed as one sample. It is expected that these samples will show the same
characteristics as any other sampling of this population or area; that is, plotting the
frequency of a characteristic versus, the number of individuals showing that the
characteristic will produce a standard, symmetrical bell-shaped curve. (See Figure 6.2.) A
characteristic that produces such a curve when measured and graphed is a normal random
variable. There are many similar curves with different absolute shapes that are considered
bell-shaped curves. The shape of the curve and the fact that the mean and median are the
same shows that its probability distribution is a normal distribution.
This same standard distribution argument can be made for other samples. A series of
air, water, or regolith material samples are all expected