General Note
Sampling as a first but also a major activity in any analytical procedure,
exhibits many pitfalls that need to be avoided. Further on, sampling needs to
be designed according to the basic research questions the corresponding
analyses wants to answer.
2.2 Sample Strategies
Taking one sample is already a sensitive task but taking a whole sample set in a
proper and sufficient way is an intricate assignment. The main challenge is to fit the
scientific research question with the sampling design in order to obtain at the end
fully reliable and representative data. Since objectives of research studies vary to a
high extend, there exist no general procedure or strategy for the design of a sampling
campaign. Each campaign needs to be planned individually. There are some substantial parameters that need to be adapted: the sample size, representativity of
sample locations, the sample grid or density of sampling locations and the sampling
frequency.
However, based on a few examples some principal problems and the challenges
behind are introduced in the following to illustrate the complicity of designing
sampling campaigns. A first, very simple subject is the question about the amount.
As already mentioned, there are some preconditions that the sample amount should
fulfill, inter alia to be representative. one needs to decide about the sample size.
Sample size can be also interpreted as aliquotation, since sampling only takes a part
of the whole system investigated (e.g. a soil, sediments or coals). One main criterion
for defining appropriate sample sizes or aliquotation rates is homogeneity. The more
heterogeneous the matrix, the higher the variance by decreasing sample size. Or the
higher the aliquotation rate, the lower the representativity in more heterogeneous
systems. This relationship is illustrated in Fig. 2.5.
Sampling in extensive systems (e.g. coal seams, soil systems, coastal bay areas,
river catchments, . . . .), especially with higher dynamics, requires an accurate decision on the position and the density of sampling locations. The following example
will focus on the positions and partly on the density of sampling locations. If a river
water section will be studied to trace the impact of a point source emission
(e.g. effluents from a sewage treatment plant) one has to think about a suitable
position of a reference site upstream of the emission point. Then the mixing of the
contamination plume along the flow direction needs to be addressed, not only along
the river course but also with water depth. A schematic suggestion for such research
objective is given in Fig. 2.6.
A similar example but for a more static system is given in Table 2.1. If it is
intended to trace a contamination in a soil system, the depth of sampling depends on
the interaction and pathway of the pollutants that have to be studied. Only the top
layers of a soil need to be sampled and analyzed, if erosion assisted distribution and
2.2 Sample Strategies
11
Sampling as a first but also a major activity in any analytical procedure,
exhibits many pitfalls that need to be avoided. Further on, sampling needs to
be designed according to the basic research questions the corresponding
analyses wants to answer.
2.2 Sample Strategies
Taking one sample is already a sensitive task but taking a whole sample set in a
proper and sufficient way is an intricate assignment. The main challenge is to fit the
scientific research question with the sampling design in order to obtain at the end
fully reliable and representative data. Since objectives of research studies vary to a
high extend, there exist no general procedure or strategy for the design of a sampling
campaign. Each campaign needs to be planned individually. There are some substantial parameters that need to be adapted: the sample size, representativity of
sample locations, the sample grid or density of sampling locations and the sampling
frequency.
However, based on a few examples some principal problems and the challenges
behind are introduced in the following to illustrate the complicity of designing
sampling campaigns. A first, very simple subject is the question about the amount.
As already mentioned, there are some preconditions that the sample amount should
fulfill, inter alia to be representative. one needs to decide about the sample size.
Sample size can be also interpreted as aliquotation, since sampling only takes a part
of the whole system investigated (e.g. a soil, sediments or coals). One main criterion
for defining appropriate sample sizes or aliquotation rates is homogeneity. The more
heterogeneous the matrix, the higher the variance by decreasing sample size. Or the
higher the aliquotation rate, the lower the representativity in more heterogeneous
systems. This relationship is illustrated in Fig. 2.5.
Sampling in extensive systems (e.g. coal seams, soil systems, coastal bay areas,
river catchments, . . . .), especially with higher dynamics, requires an accurate decision on the position and the density of sampling locations. The following example
will focus on the positions and partly on the density of sampling locations. If a river
water section will be studied to trace the impact of a point source emission
(e.g. effluents from a sewage treatment plant) one has to think about a suitable
position of a reference site upstream of the emission point. Then the mixing of the
contamination plume along the flow direction needs to be addressed, not only along
the river course but also with water depth. A schematic suggestion for such research
objective is given in Fig. 2.6.
A similar example but for a more static system is given in Table 2.1. If it is
intended to trace a contamination in a soil system, the depth of sampling depends on
the interaction and pathway of the pollutants that have to be studied. Only the top
layers of a soil need to be sampled and analyzed, if erosion assisted distribution and
2.2 Sample Strategies
11