organic vapor analysis. To monitor the bioremediation process, we used three types of soil
samples: 1) grab samples as typically used for ADEC regulatory purposes, 2) composite samples
in which six to eight grab samples are taken on each plot and thoroughly mixed together, and 3)
soil-sock samples to reduce variability. Each sample type is summarized below. Details are given
in Section 9 "Quality Assurance Plan" of our Demonstration Plan.
Grab samples were taken from four locations of each treatment plot at the start of the
demonstration and at the fall of the subsequent two growing seasons. Each of the four locations
was sampled at a shallow and a deeper depth. These samples were analyzed for GRO, DRO,
BTEX, and residual oil using ADEC-approved methods. These data provided little utility for
monitoring the processes.
Composite samples were taken from each treatment plot at the start of the demonstration and at
the spring and fall of the subsequent two growing seasons. The rationale for using a composite
sampling technique is to account for sampling spatial variability by taking sufficient samples in
each treatment plot so that their "mean value" (the composite) better represents the "population",
i.e., the soil in the treatment plot. A total of eight composite samples were obtained from each
treatment plot at each sample time. Each of the eight composite samples were composed of ten
random samples, taken from either a shallow or deeper depth, and thoroughly mixed together.
These samples were analyzed at CRREL.
For research-demonstration sites, we used soil-sock samples in an effort to reduce variability.
This approach is not amenable to typical site implementation. The soil-sock procedure is a
derivative of that used in litter decomposition studies. Approximately 200 samples were
randomly taken prior to seeding or fertilization and mixed by rotary mixer. These large mixed
samples, generally 10 to 20 ft3 of soil, were then apportioned into fine mesh, cylindrical, open-
topped bags (soil socks) that were buried vertically in the plots from which we had taken the
samples. Sufficient bags were buried so that a soil sock could be removed from each plot at each
sampling time and sacrificed for analysis.
Where the field conditions suggested that there were areas that were different, based on initial
chemical measurements, visual clues, or landscape position, we attempted to use statistical
blocking, so that each "distinct" area included one replication of each of the four treatments.
Samples taken for the soil socks were obtained from and returned to the same block.
Soil samples were collected using hand tools, which were decontaminated between samples. The
samples were packaged in sealed bags and placed immediately into coolers with blue ice.
3.6 Analytical Procedures
Composited samples taken from the soil socks were analyzed for petroleum by several
approaches to characterize the petroleum fractions in the soil. Total petroleum hydrocarbon
(TPH) data are expressed as a concentration of mass of petroleum per mass of soil. Although this
approach measures an integrated value of the total amount of petroleum products present, you
cannot distinguish among specific compounds, degree of weathering, or degradation in the form
in which TPH is usually expressed. We therefore used TPH in conjunction with more specific
methods to determine contaminant degradation and the time-related depletion of specific
fractions. The approaches are described below. Details of analytical methods are given in