face soils (015 cm) in a heavily blasted area in
AmDNT were detected in soils at the inland
range 44. An enormous amount of debris from use
ranges, but measurable concentrations were al-
of LAW rockets was found at this location, and it
ways low and confined to the surface soils. In most
is the LAW rocket that is the source of these HMX
instances, the concentrations of the two amino
residues. This is attributable, at least in part, to
transformation products exceeded that of TNT.
the high percentage of these rockets that do not
Since these three compounds are more sorptive to
detonate on impact, sometimes spilling the explo-
soils than is RDX, these low concentrations will
sives on the soil surface.* It is not possible to as-
persist in the surface soils longer than RDX, even
sess whether HMX concentrations have declined
though soils at the inland ranges are largely coarse
substantially in these soils since the 1994 study,
sands and have little retentive capacity.
because there were so few samples analyzed at this
The largest problem for site characterization at
location, and the exact locations of the sampling
the inland ranges is the extreme heterogeneity in
points are not known. Concentrations of HMX in
spatial concentration that is present at all distance
subsurface soils were much lower than in the sur-
scales. To demonstrate just how severe this prob-
face, but in a few locations, concentrations above
lem can be, a log-normal distribution was con-
10 mg/kg were measured in soil at the 15- to 30-
structed, based on the HMX data from the four
cm depth. In fact, at several sampling locations in
discrete samples collected from grid A, sampling
the grid next to the target, concentrations above 5
location 1-44 (Fig. 10). This figure clearly shows
mg/kg were observed in soil from the 105- to 120-
that the concentration of HMX in discrete samples
cm depth. It is uncertain whether the HMX at this
from this area can vary dramatically from sample
depth had leached in aqueous solution in down-
to sample. Providing a good estimate of the mean
ward percolating water, or had migrated through
HMX concentration for this area would require the
this coarse-grained soil as very fine particles. It is
analysis of a large number of discrete samples.
possible that the soils had been physically dis-
While this problem has been identified at other
turbed to this depth by the tremendous number
explosives-contaminated sites (Jenkins et al. 1996,
of large explosions that took place over many
1997a), the heterogeneity observed at the inland
years. It is also possible that, in the process of
ranges is the most extreme observed thus far.
obtaining these deep samples, they were contami-
Before performing any more analyses at the in-
nated by surface soils with a high HMX concen-
land ranges, our first priority must be to develop
tration.
improved sampling and sample processing pro-
Concentrations of RDX in the inland ranges are
cedures. Our results reveal that one reason for the
barely detectable in most locations. The detectable
poor precision and accuracy is the variable losses
levels are largely confined to the surface 015 cm
of fine dust that is highly enriched in HMX. Pos-
at range 44. Concentrations only exceeded the 0.5-
sible cures, such as slight elevation of soil mois-
mg/kg action level at the most heavily impacted
ture using a small garden sprayer or other mist-
area in range 44, and there only in the grid located
ing device during sampling and sample process-
within 5 m of the target. Concentrations of RDX
ing, should be investigated with existing samples.
have clearly declined in these surface soils since
Whatever method ultimately evolves, it must be
the 1994 study. Natural attenuation of RDX in the
adaptable to field use, starting with the removal
surface soils will undoubtedly continue, and, even
of soil samples from the ground. The technique of
if no remedial actions are taken, RDX will be un-
mixing in place before a sample is removed (used
detectable in surface soils within several years. Our
in this study) may have resulted in some loss of
experience tells us that RDX neither biodegrades
fines. Either this practice should be omitted or soils
nor degrades chemically under aerobic conditions
should be moistened while mixing takes place. The
(Grant et al. 1995). Thus, it is likely that the RDX
need for extensive mixing may be lessened,
has migrated with downward percolating water
though, since we will recommend the use of com-
to deep into the soil profile. Once dissolved in
posite samples containing soil taken from four or
water, RDX is known to migrate rapidly in soils.
more points.
Like RDX, the analytes TNT, 4-AmDNT, and 2-
Since much of the site is free of significant ex-
plosives residue, it is inappropriate to estimate a
mean concentration for a large area. Instead, sam-
pling should concentrate on surface soils around
*Personal communication with S. Thiboutot, Cana-
target locations, to define the boundaries of con-
dian Defence Research Establishment Valcartier,
tamination. To demonstrate this point, the concen-
Courcelette, Quebec, 1998.
21
Previous Page
