Flow and transport at the hillslope scale: New experimental evidence from whole-hillslope irrigation and excavations

Lateral subsurface flowpaths and flow velocities are poorly quantified. The controls on subsurface flow characteristics are needed for better modeling of nutrient and contaminant transport and flood prediction. We present results from a whole-hillslope investigation designed to characterize unsaturated zone subsurface flow paths and velocities. Specifically, we address the questions: What are flow velocities of lateral subsurface flow at the soil bedrock interface? What controls flowpath location and extent? We performed two irrigation experiments on a 6m (parallel to stream) by 8m (upslope) section of hillslope at the Woods and Rowe Maimai experimental hillslope on the South Island of New Zealand. Lateral flow on this hillslope is monitored at an exposed trench face. In the first experiment water with a dye was pumped into a pit excavated to bedrock 8 meters upslope the hillslope base until steady state conditions were reached at the trench face. Flowpath locations were then identified by excavating soil in 0.2-0.3 m slices upslope from the trench. A Bromide solution was injected every 0.4-0.6 m to determine flow velocities. The second experiment was identical except that water was applied on the soil surface as a line 4 m upslope of the trench. Once the hillslope was cleared of soil, the irrigation experiment was repeated on the exposed bedrock. A very high resolution DEM (grid spacing = 5mm) of bedrock topography was constructed using a Trimble laser scanner. The dominant subsurface flow paths were continuous and predominantly at the soil bedrock interface, less than 0.1 m above the bedrock in both experiments. In the pit experiment unbroken flow paths along the entire 8m excavated hillslope were observed. In the second experiment water applied to the soil surface moved via subvertical cracks through the soil profile for <1.5m and then traveled laterally along the bedrock interface. The bedrock topography proved to be the dominant control on flowpath locations. Topographic analysis of the bedrock suggests that while the correlation length scale is on the order of 2m, small scale (<0.1 m) features significantly affect flow routing, and had a large impact on final flow path location. The tracer experiments indicate that subsurface flow velocities at the soil bedrock interface were on the order of 0.25 cm/s - more than an order of magnitude faster than predicted in the Darcy flow regime (with laminar flow assumptions) but more than 2 orders of magnitude slower than predicted in the Hagen Poiseuille flow regime (with pipe flow or open channel flow assumptions). Tracer velocities were similar for both the surface and direct bedrock injections. These results suggest that new models of lateral subsurface flow paths and velocities are needed. Our measured flow velocities poorly fit standard models used to define water delivery. Our experiments show that measuring bedrock topography may improve our ability to identify subsurface flowpaths. These observations are used to develop a new conceptual model of lateral subsurface flow.