How Do Near-Surface Strength Characteristics Vary Over Landscape Scales? A Case Study Using Geotechnical Field Methods and a Back-Analysis of Earthquake Triggered Landslides.

Abstract:

Near-surface mechanical properties of rock are important to assessing landslide hazards and the erosional sculpting of landforms. Variations in bedrock strength are often invoked to explain patterns in landsliding and topography that are not associated with a clear tectonic mechanism. Yet few studies have attempted to quantitatively map near-surface mechanical rock properties over regional scales. Here we present new seismic and geotechnical characterizations of the shallow (to depths of 20 m) subsurface across a broad swath (about 60km x 140km) of the Himalayan Range front north of Kathmandu, Nepal. By pairing 86 1D shear wave velocity surveys and 117 engineering outcrop descriptions following the Geological Strength Index (GSI) classification system, we were able to constrain rock shear-strength characteristics in the shallow subsurface. Furthermore, we supplement our geotechnical field data with estimates of near-surface strength inferred from a back-analysis of stability for the landslides triggered by the 2015 Mw7.8 Gorkha Earthquake, also in central Nepal. From both the field data and the inverted landslide inventory, we find considerable variability in near-surface strength that is strongly dependent on the degree of weathering (which we quantify with the GSI system), rather than the between the crystalline metamorphic rocks along the Himalayan Range front. Furthermore, we resolve a strong topographic signal in both datasets. Bedrock at ridge top sites tends to be highly weathered and have strength characteristics typical of stiff engineering soils, whereas sites near bedrock channel bottoms tend to be unweathered and characterized by shear strength estimates typical of hard rock. Our strength estimates also depend strongly on topographic slope, reflecting the tendency for steep slopes to have a thinner mantle of weak, weathered material. These results underscore the potential for a high degree of heterogeneity in the near-surface strength profile over landscape scales. Future landslide hazard analysis could be significantly improved with a model for bedrock strength that leverages the combination of field methods and slope stability back-analysis, such as we present here.