8.22.
Source monitoring and environmental monitoring should be aimed at obtaining representative values. This means that the sample should reflect the conditions of the source or the environment from which it is taken. In general, activity levels in discharges or in the environment are subject to spatial and temporal variability, and the sampling procedures should be formulated to consider such variabilities (see, e.g., Ref. [57]).
8.23.
The sampling frequency should be established on the basis of the quantity to be measured, the precision needed, the time dependence and the variability of the quantity49. In general, sampling should be more frequent the higher the spatial and temporal variability. For example, more frequent sampling is needed for monitoring radionuclides with short half‑lives and food for which there is a short time period between harvesting and consumption.
8.24.
Various methods and statistical schemes can be used to ensure representative sampling in the environment: specific procedures are suggested in Ref. [58]. Although these procedures might not eliminate the uncertainty associated with activity levels in environmental samples, they may reduce the uncertainty and enable it to be quantified by statistical analysis. Table 3 summarizes the main sampling approaches and their features.
Sampling approach |
Description |
Comment |
Judgmental sampling |
Sample is taken based on the understanding of the environment and exposure pathways |
Increased probability of biased sampling; representativeness cannot be quantified |
Simple random sampling |
Any sample has the same probability of being included |
Provides samples that are representative of the sampling area; problems might arise if the area is not homogeneous; samples are not reproducible |
Stratified sampling |
The sampling area is divided into parts (strata) that are known to be more homogeneous; simple random sampling is then applied to the strata |
Requires knowledge of the inhomogeneity of the sampling area; might lead to bias if the strata are not properly estimated |
Systematic sampling |
Starting from a randomly selected point, sampling follows a strict predefined sampling grid |
In comparison with random sampling, easier to implement in practice; spatial pattern, spatial trends or correlation ranges of contamination data might be unnoticed |
8.25.
Sampling procedures should be developed to ensure that: each sample is representative of the sampled medium; collected samples are spatially independent; the sampling procedure is reproducible; and sample integrity is maintained. Procedures should be in place for quality assurance in sampling and the analysis of uncertainties originating from sampling in reported results (e.g. split samples, field replicates, field blanks) and for proper sample tracking through a chain of custody process. Technical considerations for sampling that might apply to facilities in planned exposure situations are presented in the Annex.