A NEW FORCED GRADIENT TRACER TEST TO ASSESS THE IMPORTANCE OF PREFERENTIAL FLOW PATHS ON SOLUTE TRANSPORT AT THE MACRODISPERSION EXPERIMENT (MADE) SITE

A new tracer experiment (MADE-5) was conducted at the well known Macrodispersion Experiment (MADE) site to investigate small-scale transport processes. The test was performed under dipole flow forced-gradient conditions and a bromide solution was injected for about 6 hours in a well screened over the entire saturated thickness of the aquifer. Concentrations were monitored in an extraction well and in two multilevel sampler wells (MLSs) located within a distance of 6 m from the source. The shape of the breakthrough curve (BTC) measured at the extraction well is strongly asymmetric showing a fast arriving peak and an extensive late-time tail. These characteristics suggest that a fraction of the bromide tracer was able to travel along preferential flow paths while most of the mass moved slowly in a relatively low-permeable matrix. This transport mechanism also explains the low magnitude of the observed peak compared to the initial injected concentration and the poor mass recovery (about 50%). The BTCs measured at seven different depths in the two MLSs are radically different in terms of shape, arrival times and magnitude of the concentration peaks. All these characteristics clearly confirm the presence of a complex network of preferential flow pathways controlling solute transport at the MADE site. Collected experimental data were also used to evaluate two transport models including a stochastic advection-dispersion model (ADM) and a dual-domain mass transfer model based on a deterministic reconstruction of the aquifer heterogeneity. For the integrated BTC at the extraction well, none of the stochastic ADM realizations were able to match the observed concentrations. On the other hand, the dual-domain mass transfer model is able to predict accurately the magnitude of the concentration peak and its arrival time (less than 1.5 % error). For the multilevel BTCs between the injection and extraction wells, neither model could reproduce the observed values, suggesting that a more detailed characterization of the aquifer heterogeneity to the decimeter or smaller scales may be needed for accurately representing 3D transport behavior.