Lay summary
From precipitation to run-off water passes through soils. It typically bends somewhere from vertical to lateral flow in hill slopes. We wanted to know where bending most significantly occurs, either in the unsaturated soil profile due to horizons or at the interface between soil and bedrock.

Field investigations
We installed 10 triplets of TDR-wave guides (rods of stain-less steel, 6 mm in diameter, 15 cm long) such that they formed the sides of a tetrahedron pointing into the soil. They reached to depths from 11 to 40 cm. Recording variations of soil moisture with Time Domain Reflectometry (TDR) is based on the variation of the di-electric number of the soil-water-air system that is converted to the content of volumetric soil moisture. Wetting fronts move down in the soil profile in response to either rain or sprinkling. Ideally, the true velocity vector of a wetting front is the summation of its three vectors that were recorded along each of the three TDR-wave guides.
Flow bending occurred indeed within the soil profile, however, velocities and flow rates along bedrock are between 10 and 100 times more intense. The approach did not produce reasonable results in about 20% of the cases due to heterogeneity of flow i.e., break-up of supposed coherent wetting fronts into fingers and similar features of preferential flow.
The remainder of the project deals with the impact of the flow heterogeneity on the TDR-signals and on modeling of preferential flow with a momentum-dissipation-approach that depends entirely on TDR-measurements.

Water flow in sand boxes recorded with neutron radiography.
The boxes, made of aluminum sheet metal, are 20 cm wide, 40 cm tall and 0.5 cm thick. They are filled with either coarse, intermediate, or fine sand. Infiltration is due to a sprinkler that is mounted on top of the box. A balance records continuously drainage flow from the tank bottom. Preliminary experiments, with glass panels replacing sheet aluminum, showed flow in the coarse sand along fingers and flow in the fine sand proceeding as coherent wetting fronts, while the intermediate sand produced intermediate features. At Paul-Scherrer-Institute the experimental device was set in front of a neutron beam, the images of the spatio-temporal distributions of gray densities were recorded and subsequently converted to distributions of water contents at the millimeter-scale. First neutron radiographs are merging. With their details we assess the relationship between the TDR-measurements and the actual heterogeneity of wetting fronts. We then will assess the impact of front heterogeneities on modeling infiltration at the scale of the sand tank model.