Page 232 - Environmental Control in Petroleum Engineering

218 Environmental Control in Petroleum Engineering

surveys to characterize the extent of brine plumes, can also be used
(Young, 1991; Dalton, 1993). A detailed discussion of sampling and
analysis is given by Johnson and James (1989).
Sampling procedures must be designed to provide unbiased data
throughout the cleanup process (American Petroleum Institute, 1983 and
1985a). Procedures that can contaminate samples include how the sampling
device is emplaced at the site. Drilling can alter the in situ geochemical
environment by flushing with drilling fluid or by allowing the commingl-
ing of fluids in different zones. The presence of foreign materials like
grout or bentonite around the wellbore can contaminate water flowing into
the well. To flush such contaminants from the well, typically eight to ten
well volumes of water are pumped from the well before samples are taken.
Samples can also become contaminated from exposure to atmospheric
pollutants, particularly oxygen. The composition of obtained samples can
be altered by degassing or by sitting stagnant for a few weeks before
testing, allowing bacteria to grow. Finally, human error in any of the stages
of obtaining the sample can result in sample contamination.
For groundwater remediation projects that are expected to take a
number of years to complete, the timing of sampling during remediation
should be systematic, not random. Systematic sampling is easier to
schedule and administer and will allow seasonal variations to be identified
and accounted for. For many cleanup projects, sampling four to six times per
year may be adequate to ensure that the cleanup dynamics are observed
and to minimize expensive redundancy (Johnson and Jennings, 1990).
Once obtained, the samples must then be accurately analyzed, A
variety of analytical procedures are available for different contaminants
(National Research Council, 1985; American Petroleum Institute, 1985g
and 1987a). For compounds with concentrations in the parts-per-
million or higher range, the accuracies of most analytical procedures
are generally good. For trace contaminants, however, consistent results
may be difficult to obtain. Different measurement methods can also
yield different results, particularly if improper extraction methods are
used (Holliday and Deuel, 1993). Sophisticated analytical techniques
like gas chromatography may be required for accurate results. Regulatory
agencies may specify the types of analyses that must be conducted.
After the data have been obtained, a statistical analysis of the data
must be conducted. The type of statistical analysis conducted will
depend on the needs of the study and how the resulting conclu-
sions will be used. A variety of statistical tools are available through
time series and trend analysis. When analyzing data, however, it is