5
1
ing
R
1
4
8
2
1
3
8
2
1
8
2
2
8 1
2
1
5m
cm
8
2
Distance Between Each Ring
Tank
7 7 7 7 7
33333
Target
6
4
5
6
4
6
4
5
6
4
For Each Ring:
5
6
4
1 3 5 7 Composite A
5
2 4 6 8 Composite B
5
Figure 12. Proposed concentric ring sampling plan for target areas.
concentrations exceed RDX by at least an order of
tativeness of each sample) by "averaging out" spa-
magnitude. This was the case in all areas where
tial heterogeneity effects, and 2) our need to mini-
the concentrations of HMX exceeded 2 mg/kg. The
mize the loss of detection capability for hot spots
EnSys method was not useful for estimating RDX
because of the dilution effects of compositing.
concentrations at the inland ranges because of its
While the detection limits for composites com-
response to HMX. If RDX and HMX are present in
pared to discrete samples are theoretically reduced
similar concentrations, this method will respond
by 1/k, where k is the number of discrete samples
to both and provide an estimate that is more
combined in a composite, such detection capabil-
heavily weighted by RDX, which has a response
ity could only be lost when one discrete sample is
factor that is about twice that of HMX.
contaminated and all others in the composite are
For example, if RDX and HMX are both present
barren. Given the method of contamination at a
at 1 mg/kg and the method is calibrated with
firing range, i.e., multiple explosions in a small
solutions containing RDX, the determined concen-
area, and the relative closeness of discrete samples
tration would be 1.5 mg/kg RDX. If, on the other
to be composited, we cannot envision the exist-
hand, the method was calibrated using solutions
ence of this extreme condition. None of the data
containing HMX, the determined concentration
we have collected for discrete samples in this study
would be 3 mg/kg HMX. Thus, if this method is
and elsewhere (Jenkins et al. 1996, 1997a) even
selected, it is important to specify the calibration
remotely approximate this condition. In fact, if the
method required. For characterization of the
distribution was actually this extreme, an entirely
inland ranges, we recommend calibrating with
different approach designed for locating hot spots
HMX. An MDL for HMX has been estimated at
would be required (Gilbert 1987).
1.6 mg/kg (Jenkins et al. 1995), but it can be low-
The two most widely used on-site methods for
ered to about 0.4 mg/kg, if necessary, by chang-
RDX were evaluated with selected soil samples
ing the soil-to-solvent ratio from 20 g/100 mL to
from the inland ranges. Results obtained with the
40 g/50 mL.
DTECH immunoassay method demonstrate that
If it is necessary to also provide on-site analysis
it will not provide useful results at this site for RDX
for TNT in soils from the inland ranges, the colo-
because of the level of interference caused by the
rimetric-based TNT test can be run with the same
presence of HMX at concentrations several orders
acetone extract used for the HMX test. This test
of magnitude above RDX. The response to HMX
was not evaluated with soils from the inland
is not sufficient to determine HMX either. There-
ranges, but data from CFB-Valcartier indicate that
fore, the DTECH RDX method is of no use in char-
it is quite acceptable for this application. The TNT
acterizing the inland ranges at Fort Ord.
test conducted with a soil-to-solvent ratio of 40
The colorimetric-based EnSys method was also
g/50 mL should provide an MDL of about 0.3
evaluated. It is a simple, reliable on-site method
mg/kg. A kinetic study conducted here indicated
for determining HMX concentrations when HMX
23
Previous Page
