Improved understanding of unstable finger formation in partially wettable soils

Understanding how water infiltrates soils is essential in groundwater recharge and surface runoff predictions and agrochemical contamination assessment. Infiltration fronts are often envisioned as flat, uniform wetting fronts in which water content and pressure decrease monotonically. However, water infiltration occurs as unstable gravity-driven flow in homogeneous coarse-textured or water-repellent soils as fingers, bypassing most of the soil. A characteristic of these fingers is that the advancing tips are near saturation, and the moisture content and pressure decrease behind the front. Despite many approaches to modeling this flow, the explanation for the increased pressure at the wetting front of these unstable flow fingers has remained elusive. We postulated a discontinuous matric potential at the wetting front resisting uniform water entry in the dry soil. Instead, water moves one pore at a time across the front, inducing high localized velocities that increase the static contact angle to a dynamic one, which the Hoffman-Jiang equation can describe. It causes the matric potential to increase at the front. These high velocities during pore invasion are akin to what is known as Haines jumps which require high-frequency sensors to detect.

In this study, we aimed to prove the hypothesis experimentally and refine the theory using a high-speed camera and high-frequency pressure measurements of water infiltration into partially wettable sand. A 30X50X1.6 mm glass flowcell was packed with air-dry quartz sand. Water infiltration to the cell was 15 μl/min. Porewater pressure was recorded at 500 Hz through a needle tensiometer at the back of the flow cell wall. A high-resolution, high-speed camera recorded the pore invasion at 500 fps over an area of 32X32 mm surrounding the tensiometer. A second set of experiments was performed at even greater magnification, looking at a single pore (5X5 mm) where the wetting front’s contact angle could be measured visually during infiltration. The results showed that water advanced as a series of invasion events through one or two pores lasting a few milliseconds separated by longer periods where the front was static. The invasion pore flow velocities exceeded the saturated hydraulic conductivity by three orders of magnitude. In addition, the high magnification experiment found that the changes in the contact angle during pore invasion and the observed pore water velocities agreed with the Hofmann Jiang predictions. Our experimental results offer new insights into how water infiltrates the soil, as studies rarely measure infiltration with such small spatial and temporal scales.