Sample volume effects on solute transport predictions

Abstract

Saturated hydraulic conductivities Ks and solute dispersivites under saturated flow conditions λ were determined in the laboratory on cores of three different volumes (92, 471, and 1770 mL) taken at two depths in different soil layers along closely spaced parallel transects. Variances of In Ks decreased and mean λ increased with increasing core volume reflecting greater within‐core and lower between‐core velocity variations for larger cores. Mean Ks values were significantly lower for the smallest core size suggesting structural disturbance. To investigate effects of sample volume on reduced time field scale transport predictions, Monte Carlo simulations were carried out. From multinormal joint probability distributions of In Ks and In λ for the two layers, mean breakthrough curves were generated for each soil layer separately and as a two‐layer system using a vertical stream tube approach with transverse mixing ignored. To evaluate effects of uncertainty in parameter estimates due to the finite sample number, a second level of Monte Carlo analysis was performed assuming normally distributed, independent errors in the estimated means and covariances of parameters for the two layers. Differences between break‐through curves predicted for each core size were relatively small and within 95% error bounds for the means.