Factors controlling debris avalanche initiation

Abstract

Heavy rainfall during a January 1983 storm triggered numerous debris avalanches and associated debris torrents in the Smith Creek basin, western Whatcom County, Washington, United States. Four classes of debris avalanches are recognized: wedges, drainage depressions, logging roads, and discontinuity surfaces. Nine different debris avalanche headscarps that are representative of these four classes are studied in detail. The geometric configuration of each headscarp, the properties of soils comprising the headscarps, and water tables based on numerical models of hillslope hydrology are used in limit equilibrium slope stability analyses to back-calculate root cohesion (Cr) values at failure. The calculated values are consistent with those reported by others. Four groups of Cr values are proposed and correlated with forest vegetation: group I, understory, Cr range from 1.6 to 2.1 kPa; group II, scrub forest, Cr range from 2.1 to 2.5 kPa; group III, mixed forest, Cr range from 2.5 to 3.0 kPa; and group IV, old growth forest, Cr greater than 3.0 kPa. Circular and noncircular failure surfaces are tested to determine the most likely failure configuration. Most computer-generated failure surfaces are consistent with those observed in situ. One debris avalanche is attributed to hydraulic erosion rather than Coulomb shear. Soil matrix suction contributes significantly to slope stability under drained, initial conditions. Loss of soil matrix suction during rainstorms and loss of root cohesion at failure cause a sharp reduction of soil shear strength. The stability of each headscarp is modeled during a moderately severe comparison storm in December 1979. Failure was probable at only two of the nine sites, and field evidence of prior failure is found at one of these. Sites with group I vegetation were also susceptible to failure during this storm. Pore-pressure increases triggered the debris avalanches; however, soil depth, soil density, the presence of smooth bedrock discontinuities, and root cohesion are shown to be important factors controlling initiation locations. The conclusions of this study are based on modeling results, which are only verified by post-failure observations at the avalanche headscarps. Direct field measurement of the hydrologic and shear strength parameters is required before the modeling results and conclusions drawn can be confirmed. Key words: slope stability, debris avalanche, hydrology, root cohesion, forested hillslope, factor of safety, Whatcom County, Washington.