Rate dependence of residual shear strength: Roles of clay fraction and uncertainties

Abstract

Rate-weakening and rate-strengthening behaviors of soil friction are crucial in controlling landslide runouts, determining whether a slope experiences catastrophic runaway sliding or prolonged slow creep. However, objective evaluations of these effects are rare in the literature, leading to hidden uncertainties in runout modeling. Here, results from ring-shear experiments conducted at sigma = 50-150 kPa and v = 0.018-44.5 mm/min using well-characterized clay-sand mixtures are presented. To move beyond subjective interpretation, the residual shear strength and its rate dependence were quantified by applying moving median windows and Monte Carlo simulations to account for experimental noise and variability. The primary objective was to establish whether statistically significant (p < 0.05) rate dependencies could be inferred using standard laboratory equipment, despite inherent uncertainties. The analysis identified a complex control of clay type (smectite vs. kaolinite) and proportion on the response: rate strengthening characterized both clay-rich and clay-poor mixtures, while rate weakening typically emerged in balanced clay-sand mixtures. Crucially, the statistical framework revealed that approximately half of these observed trends lack significance, particularly under low sigma, where instrumental noise and stochastic particle rearrangement can mask rheological signals. Because minor compositional shifts and soil heterogeneity can lead to divergent and statistically inconclusive behaviors, it is concluded that shear rate-dependent formulations in runout models must be adopted with caution. Objective protocols and advanced testing are proposed as necessary to provide the robust experimental evidence required for reliable landslide runout predictions, integrated with detailed knowledge of the material composition.