Hydrological processes—Letters. Topographic controls on subsurface storm flow at the hillslope scale for two hydrologically distinct small catchmetns

Controls on hillslope ow pathways are still not well understood, despite the many di€erent studies reported since the pioneering work of Hewlett (1961). In a wide range of recent modelling studies it has been assumed that ow directions can be related to hydraulic gradients derived from surface topography. This assumption leads to believable maps of predictions of soil saturation and is convenient because of the increasing availability of digital terrain maps (DTMs). The validity of the resulting spatial patterns throughout the watershed, however, have not been fully assessed [although see the discussion in Beven et al. (1995) and the recent studies of Nyberg (1996) and Crave and Gascuel-Odoux (1997)]. This results from the diculty of obtaining adequate spatial coverage of hydrological variables and the scale problem of relating point measurements to predictions at the DTM grid scale. Recent hillslope analyses of soil moisture distributions have identi®ed that ow paths do not necessarily converge to hollow regions (Anderson, 1982), that they may require dynamic spatial representations (Barling et al., 1994) and that their prediction using topographically derived indices have had mixed results (Burt and Butcher, 1986; Moore and Thompson, 1996). Spatial representations of topographically driven indices, therefore, need further exploration. A possible way forward is to conduct more detailed studies of the dynamics of subsurface ow at the hillslope scale. Analysis of a natural hillslope combines integrated (trench ow) and point-scale (soil moisture) measurements coupled with a detailed understanding of surface and subsurface features. Such data provide a more comprehensive assessment of the assumption, used in many modelling studies, that hillslope ow is controlled by surface topography. In this letter we address the question of whether it is appropriate to use surface topography to describe hillslope ow paths or if it is not more appropriate to use subsurface gradients relating to the soil/bedrock interface or the hydrologically impeding layer. We present examples for two hillslopes in di€erent hydrogeological±climatic settings. Analyses of the distribution of topographic index patterns derived from digital terrain analysis for both surface and subsurface topography are related to observational patterns of subsurface storm ow along arti®cial trench face sections.