Introduction to Environmental Data Acquisition - Fundamentals of Environmental Sampling and Analysis - page 4

This text will not discuss the nontechnical or legal aspects of sampling and analysis,
but the reader should be cautious about its importance in environmental data
acquisition. The legal objective can influence the sampling and analytical effort by
specifying where to sample, defining the method of sampling and analysis, adding
additional requirements to a valid technical sampling design for evidentiary
reasons, and determining whether the data are confidential (Keith, 1996). Sampling
and analytical protocols must meet legal requirements for the introduction of
evidence in a court, and the results of a technically valid sampling and analytical
scheme might not be admissible evidence in a courtroom if the legal goal is not
recognized early on in the presampling phase. A brief introduction to important
environmental regulations will be presented in Chapter 2.
Practical tips
In governmental and industrial settings, lab notebooks are the legal docu-
ments. In most of the research institutions, the rules about notebooks are
loosely defined. In any case, date and signature are part of the GLP.
Do not remove any pages and erase previous writings. Write contact
information on the cover page in case of loss. A typical life of a laboratory
notebook ranges from 10 to 25 years (Dunnivant, 2004).
1.1.2 Sampling Error vs. Analytical Error During
Data Acquisition
During data acquisition, errors can occur anytime throughout the sampling and
analytical processes—from sampling, sample preservation, sample transportation,
sample preparation, sample analysis, or data analysis (Fig. 1.1). Errors of environ-
mental data can be approximately divided into sampling error and analytical error. In
general, these errors are of two types: (1) Determinate errors (systematic errors) are
the errors that can be traced to their sources, such as improper sampling and
analytical protocols, faulty instrumentation, or mistakes by operators. Measurements
resulting from determinate error can be theoretically discarded. (2) Indeterminate
errors (random errors) are random fluctuations and cannot be identified or corrected
for. Random errors are dealt with by applying statistics to the data.
The quality of data depends on the integrity of each step shown in Figure 1.1.
Although errors are sometimes unpredictable, a general consensus is that most errors
come from the sampling process rather than sample analysis. As estimated, 90% or
more is due to sampling variability as a direct consequence of the heterogeneity of
environmental matrices. It is therefore of utmost importance that right samples are
collected to be representative of the feature(s) of the parent material being inves-
tigated. A misrepresentative sample produces misleading information. Critical
elements of a sample's representativeness may include the sample's physical dimen-
sions, its location, and the timing of collection. If representativeness cannot be
established, the quality of the chemical analysis is irrelevant (Crumbling et al., 2001).