Drainage effects on the transient, near-surface hydrologic response of a steep hillslope to rainfall: implications for slope stability, Edmonds, Washington, USA

Abstract. Shallow landslides on steep (>25°) hillsides along Puget Sound have resulted in occasional loss of life and costly damage to property during intense or prolonged rainfall. As part of a larger project to assess landslide hazards in the Seattle area, the U.S. Geological Survey instrumented two coastal bluff sites in 2001 to observe the subsurface hydrologic response to rainfall. The instrumentation at one of these sites, near Edmonds, Washington, consists of two rain gauges, two water-content probes that measure volumetric water content at eight depths between 0.2 and 2.0 m, and two tensiometer nests that measure soil-water suction at six depths ranging from 0.2 to 1.5 m. Measurements from these instruments are used to test one- and two-dimensional numerical models of infiltration and groundwater flow. Capillary-rise tests, performed in the laboratory on soil sample from the Edmonds site, are used to define the soil hydraulic properties for the wetting process. The field observations of water content and suction show an apparent effect of porosity variation with depth on the hydraulic response to rainfall. Using a range of physical properties consistent with our laboratory and field measurements, we perform sensitivity analyses to investigate the effects of variation in physical and hydraulic properties of the soil on rainfall infiltration, pore-pressure response, and, hence, slope stability. For a two-layer-system in which the hydraulic conductivity of the upper layer is at least 10 times greater than the conductivity of the lower layer, and the infiltration rate is greater than the conductivity of the lower layer, a perched water table forms above the layer boundary potentially destabilizing the upper layer of soil. Two-dimensional modeling results indicate that the addition of a simple trench drain to the same two-layer slope has differing effects on the hydraulic response depending on the initial pressure head conditions. For slope-parallel flow conditions, pressure head is significantly reduced near the drain; however, for transient, vertical infiltration in a partially saturated soil, conditions consistent with those observed during monitoring at the Edmonds site, the drain decreases the thickness of a perched water table by a small amount.