2006 Philadelphia Annual Meeting (22–25 October 2006)

Paper No. 10
Presentation Time: 4:00 PM


HARVEY, Charles F. and OATES, Peter M., Civil and Environmental Engineering, Massachusetts Institute of Technology, 15 Vassar Street, MIT Bldg. 48-309, Cambridge, MA 02139, charvey@mit.edu

Groundwater solute transport models based on advection-dispersion and mass-transfer equations can often predict spatial averages of conservative solute concentration without using explicit maps of pore structures or hydraulic conductivity fields. However, these models do not account for incomplete mixing at small scales, and characterizing concentration fluctuations at small scales is important for correctly upscaling non-linear chemical reactions. To study reactive solute transport, we conducted highly detailed lab scale experiments where the movement of colored dye tracers and colorimetric chemical reactions were imaged in both heterogeneous media with mild differences in hydraulic conductivity, where spreading is accurately predicted by a Fickian macro dispersive model, and in highly heterogeneous media where solute spreading is better described by a mass-transfer model. By analyzing the resulting sequence of solute concentration maps, we show that the conventional coupling of conservative solute transport models with chemical reaction models can over-predict product formation because these models fail to account for small-scale segregation of reactants. However, the experimental results are well modeled by using a transport model that predicts the variance in concentration as well as the mean concentration. Our experimental results show that the distributions of solute concentrations, both reactants and products, are very well approximated by beta distributions, and that these beta distributions are accurately predicted by the modeled mean and variance. Thus, reactive transport can be accurately modeled by a coupling of conventional transport and chemical equations, but only with the addition of equations that describe concentration fluctuations at small spatial scales.