Landslide Geometry Reveals its Trigger

Abstract

Electronic databases of landslides seldom include the triggering mechanisms, rendering these inventories unusable for landslide hazard modeling. We present a method for classifying the triggering mechanisms of landslides in existing inventories, thus, allowing these inventories to aid in landslide hazard modeling corresponding to the correct event chain. Our method uses various geometric characteristics of landslides as the feature space for the machine‐learning classifier random forest, resulting in accurate and robust classifications of landslide triggers. We applied the method to six landslide inventories spread over the Japanese archipelago in several different tests and training configurations to demonstrate the effectiveness of our approach. We achieved mean accuracy ranging from 67% to 92%. We also provide an illustrative example of a real‐world usage scenario for our method using an additional inventory with unknown ground truth. Furthermore, our feature importance analysis indicates that landslides having identical trigger mechanisms exhibit similar geometric properties.

Temporal Variations in Landslide Distributions Following Extreme Events: Implications for Landslide Susceptibility Modeling

“…There is also evidence to suggest that earthquakes can transiently change landslide spatial distributions, rather than just landslide rates, across short timescales. For example, not only do coseismic landslides tend to lie at higher elevations when compared to rainfall triggered landslides (e.g., Densmore & Hovius, 2000; Rana et al., 2021), but as observed following the 1999 ChiChi event, earthquakes can also transiently shift the locations of post‐earthquake rainfall‐triggered landslides to higher slope angles (C. W. Lin et al., 2006). Collectively, the above observations challenge the assumption that landslide spatial distributions, and thus landslide susceptibility, are time independent, even over short timescales.…”

Landsifier v1.0: a Python library to estimate likely triggers of mapped landslides

“…In our preliminary study (Rana et al, 2021), we introduced a method that can classify landslide triggers by only using geometric features of landslide planforms. This initial model constitutes the first method in landsifier library, and for continuity, we briefly describe it in section 3.1.…”

“…Landslides with the same trigger morphologically cluster, for example, covering narrowly the available statistical variability of hillslope angles in a study region (e.g., Jones et al, 2021) and, thus, could have characteristic shapes reflecting their triggering mechanism, for instance, by having similar area and perimeter ratio, or size (Taylor et al, 2018;Samia et al, 2017). We developed a binary classifier that groups landslides either as earthquake-triggered or rainfall-induced based on this hypothesis (Rana et al, 2021). This initial model demonstrated that the landslides with an identical trigger indeed exhibit similar geometric properties.…”

“…Our preliminary model (Rana et al, 2021) can classify landslide triggers by only using the geometric properties of landslide polygons. Here, we introduce two additional methods for landslide trigger classification.…”

Participatory Landslide Inventory (PLI): An Online Tool for the Development of a Landslide Inventory

“…A landslide inventory should consist of two main components: a spatial component to show the location coordinates on the geometry (i.e., perimeter, area, and volume) and an attribute component, which includes the date of occurrence, type of failure, and damage caused [8,[16][17][18][19][20]. In addition to this standard information, a landslide inventory may include the geology, geomorphology, damage, and costs incurred by the landslide depending on the purpose of the landslide inventory [21].…”