Statistical Sampling Strategies for Survey of Soil Contamination - Dealing with Contaminated Sites: From Theory Towards Practical Application - page 171

Fig. 4.1 (continued) Strata used to estimate the SCDF of soil contaminants in the Netherlands
( left ), and selected sample locations ( right) (details not clearly visible; both graphs are meant to
give a general insight into soil maps)
of the Netherlands (Brus et al. 2009 ). To increase the efficiency, the study area is
stratified according to soil type and land use. The two largest strata, “cultivated
sandy soils” and “cultivated marine clay soils” cover large parts of the Netherlands.
Therefore, these two strata were subdivided into four geographical substrata each,
which improved the spatial coverage of the study area. Finally, the stratum “uncul-
tivated sandy soils” is split into “sandy forest soils”, and “other uncultivated sandy
soils” (dunes, heath). The total number of strata thus obtained was 15. The num-
ber of sampling locations within a stratum is approximately proportional to its area.
The minimum number of sampling locations was two. This is to avoid problems
with estimating the sampling variance, see Section 4.2.2 .