Examining the role of throughfall patterns on subsurface stormflow generation

Abstract

The effect of throughfall input patterns on the hydrological response of forested hillslopes is not well understood. While field studies have contributed to our understanding of subsurface stormflow generation at the hillslope scale, such work is still of limited value because of the small number of places and events that have been characterized to date and the uniqueness of each study hillslope. In recent years, virtual experiments have been used to investigate the role of topography, soil depth, bedrock permeability and storm size, on the generation of lateral subsurface flow. However, these studies have generally assumed spatially uniform rainfall, and the interaction between vegetation and its effect on the spatial structure of input (canopy interception, throughfall) for hillslope hydrologic response has not yet been explored. Here we present a number of virtual experiments that explore the interplay among hydrological inputs (temporal and spatial distribution of rainfall) and hillslope properties (subsurface topography, soil depth), i.e. physical phenomena that are sources of space and/or time variability. We address specifically the relative importance of fine-scale throughfall patterns for hillslope hydrologic response. Topography and hydrologic field observations from an existing study hillslope were used to calibrate and test a 3D Richards equation-based finite element model. Throughfall patterns were based on published throughfall patterns in an even age stand of young conifers in the Pacific Northwest. These patterns were then varied across the hillslope during the virtual experiments. Our results showed that, surprisingly, the effect of spatial input variability of throughfall on lateral subsurface stormflow generation was minimal. For our tested case, the bedrock topography control on flow generation was much greater than the fine-scale spatial variability of the input. Using a spatially uniform area-averaged “throughfall” (i.e. open rainfall reduced by some assumed fraction, which is the simplest and most common form of throughfall representations) yielded minimal differences in subsurface stormflow response. Nevertheless, using open rainfall as spatially uniform input strongly overestimated lateral subsurface stormflow, and thus, the average impact of throughfall is important for input estimation at the hillslope-scale. Overall, the effects of fine-scale throughfall patterns on subsurface stormflow generation appear to be of secondary importance compared to effects of temporal distribution of rainfall, subsurface topography and variable soil depths.