Groundwater-Related Risk Assessment - Dealing with Contaminated Sites: From Theory Towards Practical Application - page 774

For a successful calculation of contaminant transport, the following is needed:

• a detailed understanding of the geometry of the subsoil;

• an understanding of the source;

• a thorough understanding of the processes included in the contaminant transport

model along with the algorithms to describe these processes in terms of hydraulic
and geochemical processes;

• an understanding of the range of applicability of the contaminant transport model;

• an understanding of the sensitivity of the input parameters;

• a proper assessment of (the most relevant) hydraulic and geochemical input

parameters;

•

an awareness of the (lack of) reliability of the model calculations.

In other words, a contaminant transport modeller must not only understand the
software package, but must have a thorough understanding of the context of the
model application (the Conceptual Model) and the scientific background of the
model and the processes involved. This requirement is challenged at a high level
these days, since many user-friendly numerical contaminant transport models are,
often freely, available on the internet.
Obviously, experience in contaminant transport modelling and the intuitive skills
of contaminant hydrogeologists support good modelling practice. Since a decent
model does not a priori give good model results, and in order to offer the possibility
of verification, the model user is obliged to justify the modelling process.
A serious mistake that must be avoided is distinguishing between validated and
non-validated models. The phrase 'the myth of the validated model' could be the
title of a suspense novel but is, in the context of this chapter, a serious warning for
model misuse, since, in fact, not models but only model applications can be val-
idated. Validated models hardly (if ever) exist (Leijnse and Hassanizadeh 1994 ).
Therefore, the term 'validated model' must not be used as a quality mark to con-
vince non-experts of the performance of complex models. Although a series of
validated model applications might increase the reliability of the model, there are
so many conditions that could differ for other model applications (e.g., aquifer type,
degree of heterogeneity, concentration range and presence of 'disturbing factors'
such as groundwater extractions or surface water bodies) that the unconditional
reliability of the transport calculations with the same model is nothing but an
illusion.
The choice of model very much depends on the purpose of the model activity, for
example, the necessity of a rough estimate of contaminant transport velocity or the
need for a more or less precise contaminant depth profile, a local or diffuse contami-
nant pattern, etc. The choice of model also depends on the type of Risk Assessment.
In a first-tier Risk Assessment, for example, a simple conservative model calculation
could be sufficient. When this exercise results in the exclusion of unacceptable risks,
the Risk Assessment is concluded and usually no further action has to be taken.
When unacceptable risks cannot be excluded, the use of a more complex model
may be appropriate in a higher tier. These models, in higher-tier Risk Assessments,
often are numerical models.