Behavior of the mass transfer coefficient during the MADE-2 experiment: new insights

Using a dual‐porosity transport model, a more complete analysis of the MADE‐2 experiment, a natural gradient tracer (tritium) test, is presented. Results show that a first‐order, mass transfer rate coefficient is scale‐dependent and decreasing with experiment duration. This is in agreement with previous studies and predictions. Factors contributing to the scale‐dependency are errors or approximations in boundary conditions, hydraulic conductivity (K) measurements and interpolations, mass transfer rate expressions and conceptual errors in model development. In order to formulate a self‐consistent, dual‐porosity model, it was necessary to assume that the injected tracer was trapped hydraulically in the vicinity of the injection site. This was accomplished by lowering all K values near the injection site by a factor of 30, while holding all other K values, boundary conditions and parameters at their measured or estimated magnitudes. Resulting simulations, using the same scale‐dependent mass transfer rate coefficient, were then able to reasonably match the movement of the center of mass, overall plume geometry and the anomalous mass recovery ratios observed at each snapshot. The dual‐porosity model is conceptually simple, relatively easy to apply mathematically and it simulates differences in advection that are probably the root cause of dispersion in natural heterogeneous sediments. Also, a small more realistic amount of local hydrodynamic dispersion is not precluded.