5. Determining extraction time
Since the minimum adequate extraction time will vary for different sites, this is
the first parameter to be determined. One subsample from each depth of the unit
with the heaviest soil or highest organic content is extracted for 3, 10, and 30 min-
utes and the extracts are analyzed by on-site methods. Select the shortest time to
produce adequate extraction of the heaviest soil for all further analyses. If different
soil types are present on the site, either horizontally or vertically, and only a few
soils require long extraction times, the samples might be split into two groups, but
this does complicate the logistics.
6. Determining depth of contamination
One subsample from each composite is analyzed by the on-site methods. The
results of these analyses will tell us how deeply each unit is contaminated by the
various explosives analytes. If analyte concentrations in the 0- to 15-cm portion of
the cores are much higher than at the lower depths, as is often the case, these sur-
face samples may be suitable for characterizing the horizontal extent of contamina-
tion within units. Use of surface samples is always best, when appropriate, since it
is the easiest and most economical way to obtain representative samples by com-
bining several aliquots of soil from within a defined area. We call such samples
"area integrated." If there is no consistent difference in the three depths, or if con-
centrations in the composites from the 15- to 31-cm depth are as high as or higher
than in the surface 15 cm, it will be necessary to collect deeper samples to docu-
ment the depth of contamination. Deeper samples make composite sampling more
time-consuming and expensive than when surface samples are used.
7. Validating on-site analyses
Aliquots of one extract from each section (or more if desired) are used for com-
prehensive laboratory analyses. When there are no local laboratories and the
extract cannot be transported, it will be necessary to do the laboratory analyses on
separate soil subsamples, obtained as described in section 4. In any case, prior ar-
rangements should be made to ensure that these results are quickly generated as
they are required to identify interferences and misidentifications that would
invalidate on-site results.
A spike recovery study should also be done for each analyte of interest on at least
one extract from each unit using the on-site methods. Although a single spike con-
centration is acceptable, a better procedure would be one spike addition and one
matrix dilution (assuming the concentration is at least twice the quantitation limit)
to more fully characterize the linearity of analyte response.
8. Validating compositing procedure
Subsamples (20 g) from all discrete samples at one selected depth are analyzed
by on-site methods. When it appears that the 0- to 5-cm samples will be used for
characterization, they should be used here. This would be a minimum of 16 samples
when there are four or fewer units. Means of the concentrations for the four dis-
crete samples used to make up a composite can be compared to the composite
concentration determined earlier. Relative differences will tend to be greater for
concentrations close to the lower quantitation limit. In general, agreement should
28
Previous Page
