Environmental Sampling Design - Fundamentals of Environmental Sampling and Analysis - page 52

random sampling locations should be generated from the entire site of the lagoon
(perhaps including samples at D, E, and F). The selection of the appropriate
sampling design (Section 3.3) is necessary in order to have data representative of the
problem being investigated.
3.3 ENVIRONMENTAL SAMPLING APPROACHES:
WHERE AND WHEN
Discussed below are several sampling approaches commonly used in environmental
data collection defining where and when to collect samples, including judgmental
sampling, simple random sampling, stratified sampling, and systematic sampling.
Judgmental sampling is a nonstatistically based approach, and the other three are
termed as probabilistic sampling. Note that a combination of these approaches can
be used for a specific project or at various stages of the same project. There are also
many other design approaches including those innovative ones that have potential
for improving the quality of more cost-effective data. These will be briefly defined in
Section 3.3.5. Details on advantages, disadvantages, example applications, and
statistical formulas are beyond the coverage of this text. Interested readers are
referred to two excellent sources by Gilbert (1987) and EPA's Guidance on Choosing
a Sampling Design for Environmental Data Collection (EPA, 2002).
3.3.1 Judgmental Sampling
Judgment sampling refers to the subjective selection of sampling locations based on
professional judgment using prior information on the sampling site, visual
inspection (e.g., leaks and discoloration), and/or personal knowledge and
experience. It is the preferred sampling approach when schedule (such as for
emergency spill response) and budget are tight. Judgment sampling is also preferred
at the early stage of site investigation or when the project objective is to just screen
an area for the presence or absence of contamination so a prompt decision can be
made whether or not a follow-up statistical sampling is needed.
Judgment sampling is the primary representative sampling approach used for
groundwater assessment in the selection of monitoring wells (EPA, 1995). This is
because monitoring wells are complex, expensive, and time-consuming to install. In
order to best determine the nature of a suspected contaminant plume, monitoring
wells need to be placed in areas most likely to intercept the plume. Locations are
selected based on the investigator's knowledge of the suspected contaminants, site
geology, and hydrology. The alternative approaches using randomization would
likely result in too many wells that will possibly miss the contaminant plume.
Random sampling and systematic sampling discussed below are generally not used
for groundwater sampling (EPA, 1995).
There are cases when judgment sampling cannot be applied. As judgment
sampling includes no randomization, it does not support any statistical interpretation
of the sampling results. This is equivalent to saying that when the level of confidence