In April 1999, a cutbank of the Brewster River failed two times. A third failure happened on July 4, 1999. The timing of the failures (~30,000 m3) is peculiar because, spring 1999 had normal to below normal precipitation. The runout of the debris traveled further than would have been expected for similar sized landslides. I will give a brief overview of the landslide and its processes.

Landslides along the Brewster River cutbank are common. On the map above you can see where I have delineated several previous landslides. The picture below was taken around 1911. The second picture was taken in 2000. You can see that slope instability is continuous in this region.

GEOLOGY AND RUNOUT ZONE

The cutbank is made up of glacial lake sediments. The sediments are mostly silt and fine sand couplets. A cap of sand and gravel is less than 2 meters thick.

Cross-section of the Jeffersonville landslides. Runout length is ~ 290 m and the drop height is 46 m. The debris averaged 0.7 m in thickness and the maximum thickness was > 4 m.

Topographic map of the debris zone. Gray shaded area represents extent of runout.

The debris was covered with numerous trees. The orientations of the trees were bimodal. The trees within 20 m of the edge had the same orientation as the edge. The trees in the center of the slide mass had the same radial orientation as the slide mass. The trees near the slide margins were pushed by the mass, while the trees in the center of the slide mass toppled in the transport direction.

The debris ran out farther than would have been expected using other similar sized landslides as a guide. A saturated basal zone with low friction caused the long runout (6 times the drop height). The basal zone was a shear zone where small intact clay clasts were aligned by the overriding debris. The debris above the basal zone consisted mostly of intact blocks of the silt and clay couplets (up to 10 m3).

The saturated basal (and presumable high pore pressure) zone could not dewater downward, because the debris was resting on clay tennis courts. The basal zone could not dewater laterally, because of the large area that the debris covered. The path of least resistance was to dewater vertically. The vertical dewatering formed mud volcanoes on the surface, some over 1-m in diameter.

INITIATION PROCESSES

The cutbank failures occurred after at least a month with no appreciable rainfall during a season of average to below average precipitation. The cutbank is a narrow ridge, so high pore pressures from up gradient streams also did not cause the immediate failure. Neither recent heavy rainfall nor prolonged rainfall caused the slope failures.

The landslide scarp has a well-defined bench at ~ 150 m as shown by the dark gray/tan contact in the photograph to the left. This bench coincides with soft sediment deformation of the silt and clay couplets. Sediment strength tests taken up the scarp suggests that the sediment bench is stronger than the overlying sediment. The bench did not fail during the landslide therefore; bank undercutting was not the immediate cause.

The summer of 1998 was the wettest on record. We believe that the low hydraulic conductivity of the silts and clays slowed the infiltration of the record precipitation of 1998. The signal from this record precipitation finally reached the bench and lubricated the failure plain six months after the cession of the record rainfall.

"CLEAN UP"

The State of Vermont decided to remove all of the landslide debris to reestablish the 100-year floodplain. The picture on the left was taken in the 1960s and the picture on the right was taken in the fall of 1999.

IMPLICATIONS

The Jeffersonville landslides have implications on the timing and the extent of landslides in glacial sediments. Hazard zones should be delineated at least 7 times the potential drop zone. Landslides can have up to a six-month gap from the time of heavy rain to the time of failure.

Back to my homepage