SOIL SAMPLING TO DEFINE THE BOUNDARIES OF SUBSURFACE ACTIVITY DOE/ERSP TECH NOTENO. 18.0 DATED: 25 June 1979 AUTHORS: B. Friesen, DRI M. G. Barnes, DRI The usual TRU subsurface sampling method has been to profile portions of the vertical interval from 0 to 120 em. Discrete 5 em samples have been taken at 0 to 5 cm and then centered on every 20 em to maximum depth. In contrast, the fission products sampling program required information on the entire 0 to 60 em profile. Samples were taken in the intervals 0 to 5 em, 5 to 10 em, 10 to 15 em, 15 to 25 em, 25 to 40 em, and 40 to 60 cm. As a result, a number of potential subsurface excision areas were identified on Irene and Pearl. Severe time constraints on soil removal dictated that the boundaries of any potential excision area be determined as quickly and accurately as possible. The method described herein was specifically designed to achieve that goal. There are two aspects of the method: first, the use of IMP sereening to speed resampling decisions; and second, the sampling method itself. IMP Screening A set of samples taken on day 1 would be prepared for counting in the usual manner the same day. The IMP detector would be used to count the samples on day 2, and the 241Am results transmitted to the EG&G scientist by telephone as soon as the results were completed. Hard copy results would also be sent as soon as transportation became available. The data were converted to TRU and collated by the DRI statistician and the ERSP Tech Advisor. The next sampling iteration could then be planned in time for a mission on day 3. This method minimized time lags, and optimized use of sampling crews. All samples with computed TRU activity exceeding 80 pCi/g were brought to the Enewetak lab for confirmation counting. Ten percent of the remaining samples were also counted in the lab for quality control purposes. The samples were counted "as is" in the lab, so all results were reported as pCi/g TRU, wet. Table B-18-1 gives the comparison of IMP with lab results for samples near 9-S-3 on Irene, counted both ways. Agreement was generally excellent; some of the few exceptions proved to be samples containing a very high-activity particle. Sampling Method The first step in the sampling process was to take soil samples for chemistry to confirm the TRU/Am ratio, which was known to change with depth on both Irene and Pearl. If the new ratio data indicated the TRU activity was actually less than 160 pCi/g for a location, it was dropped from further investigation. Since the fission products sampling identified the depth that appeared to be above criterion, subsequent sampling checked the same interval. The intervals at 5 em above and 5 ecm below these "key" intervals were also sampled, to detect changes in the depth of the contamination "pocket". Once the horizontal boundary of the "pocket" had been determined, additional profiles were sampled within the boundaries with the usual TRU method, to determine the numberof lifts required. The sampling design is more efficient than a complete grid, in the sense of requiring fewer samples to define a boundary. It also reflects the requirement that subsurface activity be expressed as 1/16 hectare averages. Figure B-18-1 is the complete design for the first three sampling iterations. However, after the first iteration, only those samples were taken which were required to bound a location showing TRU activity exceeding 160 pCi/g. B-18-1 For example: if, in the first iteration,