Predicting subsurface stormflow response of a forested hillslope – the role of connected flow paths

摘要

Rapid flow processes in connected preferential flow paths are widelyaccepted to play a key role in the rainfall–runoff response at the hillslopescale, but a quantitative description of these processes is still a majorchallenge in hydrological research. This paper investigates the approach ofincorporating preferential flow paths explicitly in a process-based modelfor modelling water flow and solute transport at a steep forested hillslope.We conceptualise preferential flow paths as spatially explicit structureswith high conductivity and low retention capacity, and evaluate simulationswith different combinations of vertical and lateral flow paths inconjunction with variable or constant soil depths against measured dischargeand tracer breakthrough.Out of 122 tested realisations, six set-ups fulfilled our selection criteriafor the water flow simulation. These set-ups successfully simulatedinfiltration, vertical and lateral subsurface flow in structures, andallowed predicting the magnitude, dynamics and water balance of thehydrological response of the hillslope during successive periods ofsteady-state sprinkling on selected plots and intermittent rainfall on theentire hillslope area. The number of equifinal model set-ups was furtherreduced by the results of solute transport simulations. Two of the sixacceptable model set-ups matched the shape of the observed breakthrough curvewell, indicating that macrodispersion induced by preferential flow wascaptured well by the topology of the preferential flow network.The configurations of successful model set-ups suggest that preferential flowrelated to connected vertical and lateral flow paths is a first-ordercontrol on the hydrology of the study hillslope, whereas spatial variabilityof soil depth is secondary especially when lateral flow paths are present.Virtual experiments for investigating hillslope controls on subsurfaceprocesses should thus consider the effect of distinctive flow paths withinthe soil mantle. The explicit representation of flow paths in a hydrologicalprocess model was found to be a suitable approach for this purpose.