STATISTICS AND SAMPLING IN TRANSURANIC STUDIES 1 77 



term to use here because it connotes results broader than those supplied in the usual 

 application of descriptive sampling. Most people think of an inventory as supplying 

 information on both quantity and location. When sampling is used, it is necessary to 

 decide which of these attributes should be emphasized. Descriptive sampling is concerned 

 with quantity. 



In Fig. 1 descriptive sampling is illustrated by suggesting a volume-estimation 

 (integration) process. For graphic purposes, accumulation in the environment has been 

 shown as a surface. In fact, differential accumulation in, say, soil is, of course, reflected in 

 changes in concentration, and a total is usually estimated by averaging. Figure 3 shows a 

 common technique in descriptive sampling, i.e., stratification (which is described further 

 later in this chapter). 



MAP OF STRATA 



Proportion 



of area 

 Stratum i" ^^^^tum 



No. 



1 

 2 

 3 

 4 

 5 



(W,) 



0.905 

 0.060 

 0.006 

 0.006 

 0.023 



1.000 



Standard 



deviation 



(S,) 



4.23 



42.68 



221.70 



719.37 



Allocation (n^) 

 Inventory Pattern Comparison 



31 (41) 

 21 (22) 

 11 (12) 

 35 (23) 



90 

 6 

 1 

 1 



25 

 25 

 25 

 25 



Not relevant here ■ 

 98 (98) 98 



100 



Fig. 3 Example showing how sampling intensity differs according to objectives. 

 Sampling for plutonium in surface soil at GMX site (area 5) at the Nevada Test Site. 



Sampling for Spatial Pattern 



When location is the major objective, the best sampling system may be quite different 

 from that prescribed in the survey-sampling textbooks for estimating a total. So far most 

 of the relevant results in this area have been produced in geology and geography and have 

 not really begun to show up in the statistical literature or the textbooks on sampling. We 

 (Eberhardt and Gilbert, 1976) have described some of the varied aspects of sampling for 

 spatial pattern relative to transuranic studies. The rather lengthy discussion following our 

 presentation in the work cited (see pp. 197-208) should be of interest in the present 

 context. A textbook on the subject is that of Agterberg (1974). 



The basic method for describing pattern is that of drawing contour lines to show 

 regions of equal concentration (isopleths), as illustrated in Fig. 1. 



A nalytical Sampling 



Cochran (1963, p. 4) gives a good description of analytical sampling: "Comparisons are 

 made between different subgroups in the population, in order to discover whether 



