CHLORIDE MIGRATION IN HETEROGENEOUS SOIL .2. STOCHASTIC MODELING

The observed statistics of chloride breakthrough presented by Sassner et al. (this issue) were compared with predictions of a stochastic-advective modeling approach. The stochastic-advective model based on the observed spatial distribution of flow rates and on transport parameters consistent with parameter values obtained from local breakthrough curves (BTCs) agreed well with the observations. Alternative models that agreed well with the local BTCs failed to predict the large-scale BTC with realistic parameter values. The results support the assumption that compared to advection variability, local dispersion within the mobile water will often have a second-order effect on field scale solute transport. The stochastic-advective model is robust with regard to the rate of mass transfer between mobile and immobile water zones. This robustness implies that order of magnitude estimates may be sufficient for providing useful predictions of both field scale solute transport and the associated prediction uncertainty. In contrast, accurate estimation of the statistics of solute advection at the scale of interest for the transport problem is necessary.