Fast semi-analytical approach to approximate plumes of dissolved redox-reactive pollutants in heterogeneous aquifers. 1. BTEX

Various numerical reactive transport models were developed in the last decade to simulate plumes of pollutants in heterogeneous aquifers. However, these models remain difficult to use for the non-specialist, and the computation times are often long. Users who need to fit several model parameters to match predictions with field data in heterogeneous aquifers may be discouraged by the time needed to run the simulations. The objective of this paper is to provide a set of approximations that allow performing almost instantaneous calculations for transport of redox-reactive pollutants, the most common examples being benzene, toluene, ethylbenzene and xylenes (BTEX). The approach relies on two major tools: (i) the use of flux tubes (FT), a variant of stream tubes that include dispersion, and (ii) sequential superposition of the reactions (Mixed Instantaneous and Kinetics Superposition Sequence (MIKSS)). The calculation of transport is uncoupled from the calculation of reactions. The superposition principle has been used previously for the analytical solution of a bimolecular reaction of an electron donor with an acceptor and is here extended to more than one dissolved electron acceptor reacting with more than one donor. The approach is furthermore improved by including limitations of the kinetic reactions according to the availability of the reactants and by combining kinetic and instantaneous reactions. The results computed with this approach are compared to three well known numerical models (RT3D, PHT3D, PHAST) for various test cases including uniform, slightly diverted or highly irregular flow fields and several reaction schemes for BTEX. The FT-MIKSS solution gives nearly the same results as the other models and proved to be very flexible. The major advantage of the FT-MIKSS solution is fast computation times that are generally 100 to 1000 times faster than other numerical models. This approach might be a useful tool during the long fitting procedure of field data, which may be followed by one single run of a classical numerical model using the best-fit parameters.