Abstract. We present an interrupted-flow centrifugation technique to characterise preferential flow in low permeability media. The method entails a minimum of three phases: centrifuge-induced flow, no flow and centrifuge-induced flow, which may be repeated several times in order to most effectively characterise multi-rate mass transfer behaviour. In addition, the method enables accurate simulation of relevant in situ total stress conditions during flow by selecting an appropriate centrifugal force. We demonstrate the utility of the technique for characterising the hydraulic properties of smectite-clay-dominated core samples. All core samples exhibited a non-Fickian tracer breakthrough (early tracer arrival), combined with a decrease in tracer concentration immediately after each period of interrupted flow. This is indicative of dual (or multi-)porosity behaviour, with solute migration predominately via advection during induced flow, and via molecular diffusion (between the preferential flow network(s) and the low hydraulic conductivity domain) during interrupted flow. Tracer breakthrough curves were simulated using a bespoke dual porosity model with excellent agreement between the data and model output (NashSutcliffe model efficiency coefficient was > 0.97 for all samples). In combination, interrupted-flow centrifuge experiments and dual porosity transport modelling are shown to be a powerful method to characterise preferential flow in low permeability media.
IntroductionIt is well known that heterogeneities, including biogenic pores/channels, desiccation cracks, fissures, fractures, nonuniform particle size distributions and inter-aggregate pores, are widespread in the subsurface and lead to a range of preferential flow phenomena (Beven and Germann, 1982;Cuthbert et al., 2013;Cuthbert and Tindimugaya, 2010;Flury et al., 1994). The coexistence of a relatively high hydraulic conductivity (K) domain(s) and an impermeable one, often termed dual porosity, results in a non-Fickian breakthrough curve. Solute transport in such systems is often characterised by an early arrival of solutes originating from the more mobile domain (macropores) and a slow approach to the final concentration caused by diffusion into the immobile domain (matrix or microporous network). When fitting breakthrough curves, therefore, it is often difficult to differentiate between contributions from the micro-and macropore transport mechanisms. As a consequence, in recent years there has been much research into the development of effective empirical and modelling techniques to characterise solute transport processes for dual porosity systems. One method investigated has been the use of interruptedflow solute-breakthrough experiments. Amongst the original work on this topic Murali and Aylmore (1980) discussed the influence of nonconstant flow on solute transport in aggregated soil. Brusseau et al. (1989) developed a flowinterruption method for use in measuring rate-controlled sorption processes in soil systems, which was subsequently applied by Koc...