boundary, would be measured with the IMP at 25 m spacing. These data were used to draw a revised, more accurate boundary of the area to be excavated. In most cases, the new boundary enclosed less area than the original estimate, but in any event it enclosed the smallest area that would require soil removal to meet the applicable criterion. Measurements were not made at spacing smaller than 25 m after the initial cleanup efforts on Sally because the boundaries based on 12.5 m measurements there were essentially the same as for 25 m data. Estimates of the total volume of soil to be removed were based on the refined cleanup boundary and the results of soil sampling. The soil data were used to determine the maximum depth of the contamination above cleanup criterion in the soil If there were insufficient subsurface data in the cleanup area from previous sampling, additional locations were sampled using the subsurface procedure (see Section 6.9). The total volume of soil to be removed was estimated by multiplying the surface area by the depth to which soil was to be excavated. When appropriate, the cleanup area was subdivided into smaller sections, each having a different depth. In these cases, the boundaries of the small sections and the excavation depth for each and the volume estimates were transmitted to the Joint Task Group (JTG). After a soil lift had been completed, the entire lifted area and a row of points outside the boundary were surveyed by the IMP at 25 m spacing. Average activity over 0.25 or 0.5 ha was estimated by using arithmetic means of adjacent data values. If the mean for any section still exceeded the criterion, the lift-remeasure process would be repeated until the applicable criterion was met. In a few instances, additional lifts were required in an area where no elevated subsurface contamination had been expected. In those cases subsurface soil data were collected before any more lifts were taken, to provide a better estimate of the maximum depth of the soil requiring removal. When the soil removal was complete for an area, an estimate was made of the total TRU activity contained in the excavated soil) The estimate was based on the depth gradient of the TRU activity determined from subsurface soil data, before and after average activity from IMP data, and JTG's report of the total volume of soil removed. Details of how the parameters describing the depth gradient were determined and the assumptions used in making total activity estimates are in Tech Note 10.0. The final set of measurements after cleanup included the lifted area that had been used for stockpiling contaminated soil. Measurements on the stockpile areas confirmed that no contaminated soils remained after the stockpile had been transported to Yvonne for disposal. These measurements were used in determining the final surface TRU isopleths in Section 7.5. 5.2.3 Subsurface Characterization and Cleanup The approach used for subsurface characterization in the beginning of the project was to take samples on a 25m or 12.5m grid in the vicinity of each area of suspected subsurface contamination. Then, if any subsurface TRU activity above acceptable levels was discovered, samples were taken on a finer spacing around the location with elevated activity to determine the boundary of unacceptable contamination. Each iteration of sampling was always on a finer mesh of the initial regular grid, and was intended to cover the region of interest. The first few sets of samples, from Irene and Pearl, were auger core samples. This method proved unsatisfactory, so a sidewall sampling method was used for the rest of the project. The data from the samples garly in the project consisted of gross alpha counts, with some laboratory analyses for 41am and 239,240py, The 241Am data were more useful in practice, so eventually all the samples were analyzed for 241m and some were analyzed for 239,240pu, These results were also used to determine a TRU to 24! Am ratio for subsurface soil. The sampling design changed as the cleanup project progressed. Various grid spacings and layouts of the samples were tried, but all tended to be inefficient because of the large number of sampling locations and iterations required to adequately define a cleanup boundary. Eventually the approach described in Tech Note 18 was incorporated and proved to be efficient with respect to samples and iterations, and also in minimizing the amount of soil removed. More details on the sampling designs and methods, sample analysis and cleanup methods can be found in Section 6.9. 143