Contaminant transport in natural systems is often nonideal – characterized by early peak arrivals and asymmetric breakthrough curves.  For structured soils, this behavior is explained by the existence of a mobile and immobile region with nonequilibrium mass transfer between the two.  This two-region model (TMR) requires two parameters in addition to velocity (V) and dispersivity (a_L): a, the first-order mass-transfer coefficient and b, the fraction of the mobile water content.  The TMR has also been successfully applied to heterogeneous media with no apparent immobile region, but for such media, the meanings of a and b are unclear.  Previously it has been theoretically shown that b is related to the sediment heterogeneity while a depends on V and inversely on a_L. This study investigates nonideal conservative solute transport through laboratory columns of nonstructured heterogeneous sediments to confirm theoretical findings and explore the physical meanings of a and b.

 

     Bromide flow-through experiments were conducted at different velocities on 22 cores of glacial outwash.  CXTFIT2.0 was used to simulate the breakthrough data and obtain best-fit parameter values.  For the three most well-sorted cores (sorting coefficient, C_S, < 3), the one-region model adequately simulated the breakthroughs.  For the other 19 cores, a and b values were quite variable.  Regression analysis was used to relate the variability to the sediment properties.

 

     b is inversely correlated to C_S: the greater the sediment heterogeneity, the smaller the b value.  b is not related to %silt or %clay indicating that it is not a direct measure of an immobile region, but rather a measure of the contrast in flow velocities through the core.  a is proportional to velocity, but a/V appears to be a property of the medium that like b, is inversely correlated to C_S.  Flow interruption experiments indicate that nonideal bromide-transport behavior is not a rate-limited diffusion process.  Rather than being a mass-transfer term, a is also related to the flow velocity contrast.  In fact, the product of b and a/V is more strongly negatively correlated to C_S than either parameter individually.  a_L is positively correlated to C_S and therefore, as theoretically proposed, a is inversely related to a_L.