Bringing Sustainable Management of Contaminated Sites into Practice – The Role of Policy and Regulations - Dealing with Contaminated Sites: From Theory Towards Practical Application - page 1035

Assessment is nowadays seen as an important management tool to make decisions
about environmental problems with many uncertainties.
Part VI in this topic includes information on the theory and application of Risk
Management procedures.

23.5.3 The Role of a Scientist in Risk Assessments

The scientific definition of risk as “a combination of the consequence of a negative
effect and the probability of that effect to occur”, and the large uncertainties in
risk estimation, both encourage the use of statistical approaches and the application
of decision rules based on the amount of uncertainty. There are several types of
uncertainty to be dealt with in a Risk Assessment. A number of authors have tried
to classify them. Wynne ( 1972 ), see also Doves and Handmer ( 1995 ) for instance
presents the following taxonomy of uncertainty:

• Risk : system behaviour is known, and outcomes can be assigned probabilistic

values.

• Uncertainty : important system parameters are known, but not the probability

distributions.

• Ignorance : What is not known is not known; and the degree increases when the

level of action or commitment-based on what we think we know increases.

• Indeterminacy: causal chains, networks or processes are open, and thus defy

prediction.

Shrader Frechette ( 1996 ) suggests a number of rules for scientists involved in
Risk Assessments. In Risk Assessment and environmental decisions four classes of
uncertainty are considered most relevant. The four classes are:

23.5.3.1 Framing Uncertainty

This type of uncertainty is related to the translation of the policy question in a sci-
entific question. Do we have to prove beyond reasonable doubt that there is a risk,
or that there is no risk? In the first case lack of information seems to promote safety,
in the latter case it increases the risk. It is necessary to use a so-called three valued
frame: “there is no risk”, or “the decision is not possible due to lack of information”,
or “there is a risk”.

23.5.3.2 Modelling Uncertainty

This type of uncertainty pertains to the realism of models, and to the question of
the reliability of model predictions. Very often models are considered to be vali-
dated or verified if the output of the model is consistent with some other model.
The only valid test is the comparison of model predictions with real world data. If
real world phenomena are successfully predicted one might gain confidence in the