LANDSLIDE INITIATION AFTER DROUGHT
Summer, 1998 was the wettest on record, followed by 6 months where precipitation was 24% less than normal with no long duration/heavy rainfall events. A 46 m high bank of the Brewster River in Jeffersonville, VT failed 3 times between April and July 1999, after the dry spell. The slide ran over a coherent bench of glacial silt ~5 m above the river and traveled ~290 m across the adjacent floodplain. The top of the bench had higher cohesions (direct shear test) (17 kPa) than below (5.6 kPa) and above the bench (8.2 kPa). The intact bench suggests bank undercutting was not the immediate initiation mechanism.
The debris (£ 4 m thick) consisted mostly of blocks (£ 10 m3) of clay/slit. Excavation of the debris during removal revealed a basal 20-cm thick saturated shear zone, consisting of rare 13 cm cohesive clay clasts in a homogenized gray silt matrix. This zone allowed debris mobilization, and was in sharp contact with the underlying, uneroded grass. Results from a Rf/F test suggest that landslide movement over the shear zone aligned and rotated the clay clasts. The shear zone dewatered through spectacular fields of mud volcanoes (0.5 to 2 m wide) on the surface of the debris. At the margin of the runout zone, houses were splashed with mud and steep snouts of fine-grained material suggested debris-flow like behavior and a finite yield strength of the flowing mass.
The Jeffersonville slides have implications for the timing and run out potential of landslides in glaciated areas. Timing of these mass movements suggests that landslide hazards may lag heavy or prolonged rainfall by months. The saturated basal shear zone allowed the slide to travel much farther than other similar-sized landslides.