[en] As part of a cooperatively sponsored HAZWRAP and Lawrence Livermore National Laboratory (LLNL) project to evaluate field-based methods to measure retardation factors, we investigated some aspects of the variability associated with measuring the concentrations of perchloroethylene (PCE) within the saturated unconsolidated alluvial sediments beneath LLNL. Using the depth-sampling technique developed at LLNL, we are describing the relationship of VOC spatial variability to various sediment characteristics such as geophysical log results, grain size distribution, cation exchange capacity, total organic carbon, permeability, bulk density, skeletal density, and porosity. Using this information an algorithm is being developed to calculate representative field-based partitioning coefficients (Ka) and retardation factors (R). A critical parameter used in the algorithm is the total mass of PCE present in a saturated sediment sample. The two major sources of sediment PCE concentration variability examined were those associated with (1) the analytical extraction methods and (2) the spatial changes in fine-grained materials within permeable and confining sediments. To evaluate desorption efficiency during headspace extraction (EPA Method 3810) for VOC analyses (EPA Method 8010), a bulk thermal desorption chamber (BTDC) was constructed and calibrated. The BTDC extraction provided a reference method against which generally used EPA extraction methods such as headspace were compared. Compared to untreated control surrogate sediment samples, BTDC extraction efficiencies of better than 94 percent were obtained. The results of PCE comparisons between adjacent sediment samples indicated that there was no significant difference between headspace extraction techniques and the BTDC. We used both headspace and BTDC analytical results to describe the vertical distribution of PCE in four adjacent boreholes. 14 refs., 6 figs., 2 tabs