Groundwater-Related Risk Assessment - Dealing with Contaminated Sites: From Theory Towards Practical Application - page 772

The success of groundwater transport estimates depends on the schematization
of irregular features in numerical models. MacQuarrie and Mayer ( 2005 ), for exam-
ple, modelled contaminant transport in fractured crystalline rock environments as
discrete fracture or dual continuum media. Several other models include preferen-
tial flow due to cracks or soil materials with a high hydraulic permeability. Zheng
and Gorelick ( 2005 ), for example, included this phenomenon in their model by
decimetre-scale preferential flowpaths. They used a hypothetical but geologically
plausible network of 10 cm wide channels of high hydraulic conductivity, as a rep-
resentation of the relative preferential flow paths. Field and Pinsky ( 2000 )useda
two-region non-equilibrium model (including partitioning of solute into mobile-
and immobile-fluid regions) for the calculation of contaminant transport in karstic
aquifers.
Moreover, the impact of the factors described in Sections 17.3.2.2 and 17.3.2.3
on contaminant transport, which are the presence of heterogeneous aquifers, surface
water bodies and anthropogenic subsurface processes and structures, is difficult but
essential to quantify. This phenomenon results in an uncertainty that is difficult to
assess, and which implies a challenge for practitioners.
Hydraulic and Geochemical Input Parameters
With regard to hydraulic conditions, a specific hydraulic gradient has to be assumed
over relatively large areas. Geochemical input parameters have to be determined for
large volumes of aquifer material, often of an irregular nature. As a consequence,
the description of both hydraulic and geochemical input parameters contributes to
uncertainties in the predicted groundwater quality. It is a well-known fact that, as
for any other model, the availability of adequate input parameters is crucial for the
quality of the model calculations. Several guidelines exist for contaminant transport
model parameter identification (e.g., Lovanh et al. 2000 ). With regard to the param-
eter identification, it is a huge advantage when the contaminant transport modeller
is familiar with the geometry of the soil and aquifer.
Interaction processes between water-bearing soils and rocks and the water itself
generate different water quality sectors, and this fact will determine the natural
evolution of every particular system. The effect of interactions with hydraulically
connected aquifers as well as surface water needs to be considered when looking
for a detailed description of heterogeneities. This factor is particularly important
when dealing with coastal aquifers, where a freshwater and saltwater interface
exists. Interactions between both types of waters and geological formations must be
accounted for, since seawater intrusion is a dynamic and three-dimensional process
that causes water quality variations at horizontal and vertical scales.
The processes to which calculated contaminant profiles with depth or con-
taminant breakthrough curves are most sensitive depends on the contaminant
characteristics. Mobile contaminants, for example, are generally most sensitive to
water flow-related processes and input parameters, while immobile contaminants
are more sensitive to adsorption and adsorption-related input parameters (Swartjes
et al. 1993 ).