Probabilistic Modeling of Landslide Initiation and Runout Mapping under Current and Future Climates

Abstract:

The intersection of landslides with human development creates risks of property damage, disruption of infrastructure, injury, and loss of life. Quantifying these risks is difficult due to the spatial and temporal heterogeneity of hydrologic and landscape characteristics impacting slope instability, uncertainty in landslide runout paths, and ambiguity in how climate change may impact landslide risk. Although recent progress has been made in developing probabilistic models of landslide initiation and rules for landslide runout, a fully probabilistic approach synthesising initiation and runout risks for large spatial scales has not been developed. The aim of this study is to probabilistically model precipitation-induced landslides and runout under current and future climate scenarios in the Colorado Front Range. We use an infinite-slope stability model, using the Python-based Landlab toolkit, to perform Monte Carlo simulations of probabilistic landslide initiation. The landslide stability model is coupled hydrologically with the EMT+VS (Equilibrium Moisture from Topography, Vegetation, and Soil) model, which downscales coarse-resolution soil moisture by incorporating the dependence of soil moisture on topographic and vegetation characteristics. The probabilistic initiation maps are tied to a new model of landslide runout, relating empirical observations of landslide tracks to probabilistic criteria for landslide cessation and deposition. The combined model ultimately produces estimates of landslide hazard probability. We test the model by simulating conditions over a 1333 km2 area of the Colorado Front Range for a 2013 storm that produced over 1,300 mapped landslides and debris flows. Potential changes in landslide hazard due to climate change is considered by adjusting vegetation patterns associated with potential climate change scenarios. This approach provides process-based, quantitative estimates of landslide hazard probability under current and future climates.