hazard exists, a biased sampling approach must
for only one analyte or class of analyte. Without
be adopted (Sisk 1992). Soils suspected of having
other supporting knowledge, concluding that a
high concentrations of explosives should be grab-
soil is not reactive based upon just one analysis
sampled and analyzed to determine whether the
could be dangerous. For assessing reactivity when
level of explosives exceeds 10%. Samples to be
multiple compounds are present at high levels, the
shipped for off-site analysis must be subsampled
CRREL and EnSys RISc colorimetric methods for
and analyzed on site. Explosives residues are usu-
TNT and RDX are more appropriate than immu-
ally concentrated in the top 5 to 10 cm of soil; there-
noassay test kits because colorimetric tests detect
fore, deep samples must not be collected, blended,
a broader range of explosives analytes. Some con-
and analyzed to determine reactivity. Vertical
servatism in evaluating potential reactivity using
compositing of surficial soils with high levels of
colorimetric methods is appropriate. For example,
explosives with deeper, relatively clean material
Jenkins et al. (1996c) recommended using a limit
provides a false indication of reactivity. Soils con-
of 7% explosives for conservatively estimating the
taining explosives residues over the 10% level can,
lower limit of potential reactivity. High levels of
using proper precautions, be blended with cleaner
explosives in soils may result in a low bias for on-
material to reduce the reactivity hazard and per-
site methods because of low extraction efficien-
mit shipment to an off-site laboratory, but the di-
cies. Colorimetric tests of chemical composition
lution factor must be provided with the sample. If
are used only to estimate potential reactivity. There
analytical results indicate that explosives are
are no on-site methods available to actually deter-
present at a concentration of 10% or greater, the
mine explosive reactivity. Explosive reactivity is a
samples must be packaged and shipped in accor-
determination made from validated laboratory
dance with applicable Department of Transporta-
analyses.
tion and EPA regulations for reactive hazardous
waste and Class A explosives (AEC 1994) to a labo-
ratory capable of handling reactive materials.
PROCEDURES FOR STATISTICALLY
In addition to the above information, the Army
COMPARING ON-SITE AND
Environmental Center requires certain minimum
REFERENCE ANALYTICAL METHODS
safety precautions, as summarized below, for field
sampling work at sites with unknown or greater
When on-site methods are used, their perfor-
than 10% by weight of secondary explosives
mance needs to be evaluated; this is commonly
contamination (AEC 1994). An extensive records
done by analyzing splits of some soil samples by
search and historical documentation review must
both the on-site method and a reference method
be conducted regarding the contaminated area to
(commonly Method 8330). The performance of the
identify the specific explosives present, determine
on-site method is then statistically compared to
how the area became contaminated, estimate the
the reference method using a variety of methods,
extent of contamination, and determine the period
depending upon the objective and the character-
of use. Personnel responsible for taking, packag-
istics of the data. In most cases, measures of preci-
ing, shipping, and analyzing samples must be
sion and bias are determined. Precision refers to
knowledgeable and experienced in working with
the agreement among a set of replicate measure-
explosives. Soil samples must be taken using
ments and is commonly reported as the RSD (stan-
nonsparking tools; wetting the sampling area with
dard deviation divided by the mean and expressed
water may be necessary. If plastic equipment is
as a percent), the coefficient of variation (standard
used, it must be conductive and grounded. Sample
deviation divided by the mean), or the relative
containers must be chemically compatible with the
percent difference. Bias refers to systematic devia-
specific explosive, and screw tops are prohibited.
tion from the true value.
Samples are to be field screened for explosives if
The following discussion of statistical methods
possible. Sufficient soil samples must be collected
applies to comparisons of analytical results based
to characterize the site in a three-dimensional ba-
on paired sample data, e.g., soil samples are ana-
sis in terms of percent secondary explosives con-
lyzed by both an on-site method and a reference
tamination with particular attention paid to iden-
method, or soil extracts are analyzed by two dif-
tifying hot spots, chunks of explosives, layers of
ferent on-site methods. Care must be taken in in-
explosives, discolorations of the soil, etc.
terpreting the result. For example, if split
In screening samples for reactivity, it should be
subsamples from a soil sample are analyzed by
remembered that most screening procedures test
on-site and reference methods, the differences de-
12
Previous Page
