Closing the Loop of the Soil Water Retention Curve

The soil water retention curve (SWRC) is a fundamental constitutive relation in modern soil physics, hydrology, and soil mechanics. It defines the energy equilibrium conditions between soil water content and its prevailing water potential. Fundamentally, the SWRC is governed by two groups of soil properties: pore-size distribution and interfacial physical properties. Like other constitutive behaviors of soil, the SWRC is hysteretic. Thus, depending on the initial water content and wetting/drying trajectory, the SWRC can follow paths other than the principal drying or principal wetting curves. A dominant phenomenon of SWRC hysteresis is that the water content at zero matric suction can differ significantly between the two principal SWRCs; as much as 0.20 between the two principal SWRCs has been reported (Lu et al. 2013).

Nearly all published SWRC models and test data do not extend below zero matric suction. To date, the loop of SWRC has not been considered closed, i.e., there is no information regarding the path of matric suction as a function of soil water content between these two water contents at zero matric suction. By physical reasoning, some path, likely in the negative matric suction or positive pore-water pressure regime, would close the SWRC loop. Although such a path represents continued wetting below zero matric suction, the soil is still not fully saturated.

For the first time, the authors produced the complete two principal SWRCs under laboratory conditions. An innovative testing technique combining the transient water release and imbibition method (TRIM) (Wayllace and Lu 2012) and CFM (Lu et al. 2006) was used to identify the principal paths of SWRC in the positive pore-water pressure regime under unsaturated conditions. Fig. 1 shows the experimental results of the complete loop of the SWRC for a silty soil. The testing procedure consists of two sequential stages: TRIM for the drying and wetting, SWRC for matric suction greater than or equal to zero, and constant flow method (CFM) for matric suction less than zero. The procedure described by Wayllace and Lu (2012) was followed for the TRIM testing stage. The resulting wetting and drying, SWRCs are shown in the upper part of Fig. 1. For this soil, the difference in the volumetric water content between the principal drying and wetting SWRCs at zero matric suction is $\sim 0.032$. For the CFM testing stage, several increasing constant injection rates were sequentially applied until full saturation of the specimen was achieved, and the SWRC loop is closed (five injection steps were needed for this soil). The path is shown in the inset in Fig. 1. A negative matric suction of 9.8 kPa is needed to reach full saturation for this soil, which is a sufficient change in matric suction to trigger landslides under rainfall infiltration conditions.

This work pushes the understanding of the interaction of soil and water into new territory by quantifying the boundaries of the SWRC over the entire suction domain, including both wetting and drying conditions, and thus closing the loop of the SWRC. The technique also provides a means to scan through the entire positive pore pressure domain using cyclic constant-flow rates.