Vectors of subsurface stormflow in a layered hillslope during runoff initiation

Abstract. The focus is the experimental assessment of in-situ flow vectors in a hillslope soil. We selected a 100 m² trenched hillslope study site. During prescribed sprinkling an obliquely installed TDR wave-guide provides for the velocity of the wetting front in its direction. A triplet of wave-guides mounted along the sides of an hypothetical tetrahedron, with its peak pointing down, produces a three-dimensional vector of the wetting front. The method is based on the passing of wetting fronts. We analysed 34 vectors along the hillslope at distributed locations and at soil depths from 11 cm (representing top soil) to 40 cm (close to bedrock interface). The mean values resulted as follows v_x=16.1 mm min^-1, v_y=-0.2 mm min^-1, and v_z=11.9 mm min^-1. The velocity vectors of the wetting fronts were generally gravity dominated and downslope orientated. Downslope direction (x-axis) dominated close to bedrock, whereas no preference between vertical and downslope direction was found in vectors close to the surface. The velocities along the contours (y-axis) varied widely. The Kruskal-Wallis tests indicated that the different upslope sprinkling areas had no influence on the orientation of the vectors. Vectors of volume flux density were also calculated for each triplet. The lateral velocities of the vector approach are compared with subsurface stromflow collected at the downhill end of the slope. Velocities were 25-140 times slower than lateral saturated tracer movements on top of the bedrock. Beside other points, we conclude that this method is restricted to non-complex substrate (skeleton or portion of big stones).