Evaluation of double sampling for estimating plutonium inventory in surface soil

The Nevada Applied Ecology Group (NAEG) is currently obtaining estimates of the total amount (inventory) of 239-240Pu in surface soil using stratified random sampling at a number of ''safety-shot'' locations on the Nevada Test Site (NTS) and the Tonopah Test Range (TTR). Since radiochemical analyses for plutonium are costly and large numbers of samples are usually required, the usefulness of double sampling in combination with stratified random sampling for estimating plutonium inventory was investigated. A less costly method of analysis is used to estimate plutonium concentrations using a linear regression relationship between the two variables. The regressions of 239-240Pu concentrations on the following auxiliary variables are considered: (1) Ge(Li) scans for 241Am on soil samples brought to the laboratory, and (2) FIDLER (Field Instrument for the Determination of Low Energy Radiation) net 60 keV counts per minute (cpm) readings of 241Am taken in the field. Data from two NTS safety-shot sites (Project 57 in Area 13 and GMX site in Area 5) indicate that except for the lower plutonium concentration zones, the use of either Ge(Li) scans or FIDLER counts in conjunction with 239-240Pu radiochemical analyses can yield estimates of inventory with greater precision than possible with plutonium analyses alone. Percentage reductions in standard errors up to 35 percent using Ge(Li) scans and up to 45 percent using FIDLER counts appear to be possible with no increase in total costs. Also, the data suggest that double sampling will supply standard errors equal to those being obtained using only plutonium analyses at cost savings of 20 to 30 percent. (U.S.)