Characteristic landslide distributions: An investigation of landscape controls on landslide size

Estimation of landslide erosion rates and hazard prediction require a firm understanding of the physical controls on landslide size. In this study we seek to understand how the characteristics of different landscapes and forcing events influence the distribution of landslide size at a regional scale. We explore the parameter space of a mechanically-based landslide model through a series of simulations using digital elevation data. Consistent with previous studies, we find that large slope failures are infrequent due to the scarcity of large, steep hillslopes in typical mountainous topography. On the other hand, we find that the occurrence of small landslides is limited by the cohesive strength of hillslope material, which overcomes the weaker driving forces on short slopes. We test our model results with an empirical investigation of frequency-size distributions for eight real co-seismic landslide inventories. Although empirical data are noisy, we find a positive correlation between landslide size and hillslope relief, while the effects of PGA on landslide size are less pronounced. We conclude that landslide size distributions reflect the available distribution of hillslope geometries in a given landscape, and that external forcing (e.g., seismic ground motion) determines which subset of the hillslope distribution fails during a particular event. For a given landscape and forcing event, there is a particular length-scale over which landsliding is statistically favored, leading to the concept of characteristic landslide size distributions. (c) 2020 Elsevier B.V. All rights reserved.