Macropore–Matrix Water Flow Interaction around a Vertical Macropore Embedded in Fine Sand—Laboratory Investigations

Abstract

Infiltration experiments were conducted in a large‐scale container setup to investigate water flow through a macropore and between the macropore and the surrounding porous media matrix. The experimental container had a height of 120 cm. Its base was a half disk with a diameter of 100 cm. The container was built from stainless steel with an observation glass window as a section plane. It was homogeneously filled with fine sand. An artificial macropore was placed vertically along the centerline of the glass pane. Tensiometers embedded throughout allowed for a continuous observation of pore water pressure heads. During the injection of water into the macropore, interval photography was employed to visualize the propagation of the water front in the matrix. It was shown that water is being imbibed into the surrounding matrix due to capillary forces, thus depleting the discharge in the macropore. For initially dry sand, this matrix–macropore interaction effectively limits the penetration depth of water inside the macropore. Only after long injection periods (>8 h for low injection rates) with a concurrent increase in water saturation in the porous matrix around the macropore does the vertical macropore flow increase successively and thus a penetration to deeper zones is detected. For all injection rates the water distribution in the matrix was observed visually and showed that for small and medium flow rates the lateral distribution in the matrix temporarily predominated the vertical inflow in the macropore. The experimental design allowed for the calculation of transfer velocity in water flows between the macropore and the surrounding matrix. Furthermore the transfer velocities could be related to the water pressure head in the surrounding fine‐grained sand matrix. Here a distinct decrease of transfer velocity from dry to wet conditions was observed. The transfer velocities obtained were discussed and compared with the results from similar experiments conducted by other authors. These experiments provide a large data set and thus a sound basis for testing and improving numerical models.