Basic HTML Version
www.geotechnicalnews.com
Geotechnical News • June 2013
37
GROUNDWATER
able mechanisms, including that of
the “lucky” boreholes that had hit a
subvertical fracture in the till layer.
When the author was asked to exam-
ine the data, at the beginning he was
as puzzled as the consultant. Then he
asked to see documents that usu-
ally are not given to the client, the
handwritten field reports by the field
inspectors (geologists or engineers).
The consultant was unwilling to
provide them. His position was that
the expert (the author) had already
received everything he needed, and
that the expert would lose and make
lose time and money at collecting
insignificant flaws in handwritten
documents. However, according to
the law of engineers, the consultant
must carefully keep and file the field
reports. The borehole logs were
prepared using these field reports and
the laboratory reports, but the logs
summarize information, they do not
give all information. Finally, copies
of the handwritten field reports were
provided.
As it happens frequently for ground-
water investigations, the field reports
were largely less detailed, less com-
plete than those that are required for
geotechnical investigation of future
dams or other facilities with major
safety issues. However, the field
reports gave the times at which the
drilling and sampling operations took
place. This added information was
very useful for the case history.
Using the time information, the author
found that if a till sample was highly
polluted, it was always the first sample
collected in the morning. If a till sam-
ple was moderately polluted, it was
always the first sample collected in the
afternoon. In addition, when the pol-
luted till samples were found, the last
(deepest) sample of sand–and–gravel
was also highly polluted. This finding
supported the idea that free–product
pools filled depressions of the uneven
till surface.
As a result, there was a simple expla-
nation to having found erratic levels of
DNAPL in the till layer. The DNAPL
in the upper aquifer moved downward
in the small space between the casing
and the till wall: it reached the bottom
of the hole to contaminate the till at
the toe of the casing. The split–spoon
sample contained a polluted till if
the DNAPL had time enough to seep
along the casing. All the highly pol-
luted samples were first ones in the
morning, after a night rest for the drill.
All the moderately polluted samples
were first ones in the afternoon, after
a small rest for the drill during the
lunch. Clearly, the DNAPL contamina-
tion of the till aquitard was cross–con-
tamination during drilling, and mostly
between drilling periods. During the
morning and the afternoon, it seems
that the DNAPL had no time enough
to move along the casing to reach the
next sampling level.
After finding the reason for having
erratic DNAPL levels in the till layer,
the author (expert) recommended
additional investigation. Surprisingly,
no geophysical surveys had been
done for this case, whereas the simple
stratigraphy was perfect to obtain clear
geophysical information. The subse-
quent geophysical surveys revealed
that the aquitard (till layer) was
discontinuous, with large “windows”
between the sand–and–gravel aquifer
and the fractured rock aquifer. This
was afterwards verified using bore-
holes that confirmed the till absence
and thus the till layer discontinuity.
Therefore, the till aquitard could
not protect the underlying fractured
rock aquifer against fast and intense
contamination. The windows easily
explained the high level of pollution in
the rock aquifer.
Conclusion
DNAPL cross–contamination may
happen from an upper unconfined
aquifer down to a confined aquifer, via
natural paths such as large apertures
(windows, pinchouts) and small
apertures (near–vertical fractures) in
the aquitard, and via man–made paths
such as vertical spaces along drilling
casings, incorrectly sealed monitor-
ing wells, and long–screened MWs.
All these potential pathways must be
investigated.
When a borehole is drilled, or when a
penetrometer is driven, through a con-
taminated unconfined aquifer, there
is a small space between the casing
and the hole wall. This small annular
space acts as a conduit or preferential
pathway for any DNAPL that can
travel downwards along the casing
and reach a deeper aquifer. A drilling
method using a dense mud may help
to prevent this migration; however the
use of mud may cause further prob-
lems, especially for developing the
MWs and then collecting representa-
tive groundwater samples.
In the worst scenario, the DNAPL
reaches the aquitard material as soon
as it is sampled, or after a few hours.
This is what happened in the 1980s
case of this paper. Similarly, U.S. EPA
(1992, 1994) warned groundwater pro-
fessionals that DNAPL can be carried
to greater depths during drilling. This
drilling cross–contamination creates
a bias in soil– and water–quality sam-
ples, leading to an incorrect evaluation
of groundwater contamination.
Cross-contamination may also happen
after drilling due to a poor sealing of
the borehole when installing a MW, or
due to non–sealable damage or defects
in the wall, that were created by
drilling (Chapuis and Sabourin 1989;
Meiri 1989; Avci 1992; Lacombe et al.
1995; Lapham et al. 1995; Yesiller et
al. 1997; Chapuis 1998; Chapuis and
Chenaf 1998; Chesnaux et al, 2006;
Chesnaux and Chapuis 2007), or seal
deterioration (U.S. EPA 1992, 1994).
It may also be due to long–screened
monitoring wells, which act as long
man–made vertical drainage paths
(Church and Granato 1996; Santi et al.
2006; Mayo 2010). From an engi-
neering point of view, wells may be
viewed as a weakness in a natural bar-
rier system (Warner 1996). Wells may
be designed, drilled, and constructed
to try to minimize cross–contamina-