Recent advances in dynamical models of bioremediation

A cellular automata model of the complex set of interacting phenomena which take place in bioremediation has proven able to accurately describe several experimental data on a pilot plant. There are some parameters in the model which are chosen in such a way as to match experimental data. In this paper it is shown that the model has interesting generalization capabilities, i.e. that it can lead to accurate predictions also in some cases which have not been used for parameter adaptation.Moreover, some possible variants are examined : the first one amounts to a model simplification by resorting to a local equilibrium assumption for the adsorption/desorption process, while in the second one a cellular automata with time variable step is introduced, in order to avoid some nonphysical situations which might arise in the original model. Both variants are tested and discussed, and it is shown that they do not lead to improved performance, at least on the available set of experimental data.