Introduction to Contaminated Site Management - Dealing with Contaminated Sites: From Theory Towards Practical Application - page 35

in the pore water, the metal concentration adsorbed to the root surface, the metal
contents in the roots and the metal contents in the shoots.
In Hodson et al. ( Chapter 16 of this topic) a detailed description of bioavailability
is given.
1.5.4 Reliability
1.5.4.1 Uncertainties and Variability
It is generally acknowledged that Risk Assessment, although it is said to be an
objective process, is also an unreliable process (e.g., Ferguson et al. 1998 ). There
are several reasons for this. First, Risk Assessment includes many parameters and
equations that have large uncertainties and variability. Uncertainty is the variation
in these Risk Assessment tools due to lack of knowledge or to lack of scientific con-
sensus. Variability is the variation due to spatial and temporal variations. The large
variability is explained by the heterogeneous nature of soil and the large differences
in human characteristics and behaviour among individuals. The large uncertainties,
often found in exact sciences, are mostly related to the transfer of contaminants from
soil into contact media, and biokinetic fate and transport processes in the human
body (Human Health Risk Assessment and Food Safety), the immensely complex
functioning of the soil ecosystem with mutual interactions between many differ-
ent organisms and the soil properties (Ecological Risk Assessment), and transport
processes in soil and aquifers (Groundwater-related Risk Assessment). Every vari-
ability found in the Risk Assessment tools is interwoven with uncertainty, while
uncertainty does not necessarily go together with variability. Examples of equa-
tions that are characterised by both large uncertainties and variability are transport
processes of contaminants through the aquifer, pore water and soil gas, and the
equations that describe bioavailability in soil. A parameter that mainly has large
variability is, for example, the fraction of total vegetable consumption that humans
grow in their own garden. For a specific case this fraction can be accurately esti-
mated, but for the use of a generic value for the derivation of Soil Quality Standards
there is a huge variation between sites, and most definitely for bigger geograph-
ical entities. Generally speaking, patterns that describe human behaviour, and the
behaviour and composition of soil ecosystems show a wide variation in time and
space.
Second, Risk Assessment includes a whole chain of calculations and measure-
ments, which means that small uncertainties in an earlier step (e.g., in the sampling
strategy) might add up to large uncertainties in the final step (e.g., the risk character-
isation). Third, several elements in Risk Assessment require a subjective judgment,
which means that quite a number of uncertainties are involved with the sometimes
arbitrary choices of the risk assessor.
It must be realised that measurements, although the general belief is that these
are much more accurate than calculations, are also often characterised by lim-
ited reliability. Nevertheless, in some specific cases, the reliability could indeed be
improved by including measurements in the calculations, namely, measurements of