Our results also provide a physically based and experimentally verifiable explanation for landslide creep. In short, the inherently rate strengthening behavior of clay-rich shear zones may explain the behavior of slow landslides developed in Franciscan mélange.

This, in turn, suggests that it may be possible to predict which landslides are prone to catastrophic acceleration on the basis of paired geologic mapping and laboratory measurements of frictional properties. For example, creeping landslides revealed by synthetic aperture radar interferometry (InSAR) in the western United States are commonly associated with surface exposures of clay-rich subduction mélange.

Where these rocks transition to quartz-rich turbiditic sandstones to the north, evidence for creeping landslides largely disappears, although evidence for landsliding is still ubiquitous in the topography.

This potential lithologic control on the style of landslide failure is notable and raises the possibility that fundamental differences in landslide behavior—and hazard—in these two settings, and elsewhere, could be controlled by opposing frictional properties (i.e., rate strengthening versus rate weakening) in the contrasting lithologies in the two settings, a hypothesis that is eminently testable through targeted experiments of rock friction.