English Abstract Mathematical models are frequently used in risk
analysis for underground disposal of nuclear waste. The validity of these
models is examined in an international validation project INTRAVAL. RIVM
participates in INTRAVAL project for the validation of the code METROPOL, a
groundwater flow and solute transport program developed at RIVM. Within
INTRAVAL several test-cases have been formulated. One of these test cases
is an in-situ migration experiment in an argillaceous formation at 220
meters depth at Mol, Belgium. This experiment has been set up to examine
whether parameters derived from laboratory experiments may be used for long
term predictions on a larger scale. The present report describes the
simulations of the experiment with the METROPOL code and it discusses an
analytical solution of the problem. The simulation results indicate that
the model concept of the METROPOL code provides an adequate description of
the tracer migration in the Boom clay at the Mol site. However, to prevent
discretization errors, a fine mesh is required to model the process in areas
where steep concentration gradients are expected. Application of fine
meshes leads to a larger number of nodal points and, consequently, to an
increase of required computer memory and CPU time consumption. With respect
to the parameters from the laboratory tests, it is shown these can be used
for long term predictions of problems at field scale.
