Frequency-area distribution of earthquake-induced landslides : abstract

H. Tanyas, Kate E. Allstadt, C.J. van Westen

Research output: Contribution to conferenceAbstractOther research output


Discovering the physical explanations behind the power-law distribution of landslides can provide valuable information to quantify triggered landslide events and as a consequence to understand the relation between landslide causes and impacts in terms of environmental settings of landslide affected area. In previous studies, the probability of landslide size was utilized for this quantification and the developed parameter was called a landslide magnitude (mL). The frequency-area distributions (FADs) of several landslide inventories were modelled and theoretical curves were established to identify the mL for any landslide inventory. In the observed landslide inventories, a divergence from the power-law distribution was recognized for the small landslides, referred to as the rollover, and this feature was taken into account in the established model. However, these analyses are based on a relatively limited number of inventories, each with a different triggering mechanism. Existing definition of the mL include some subjectivity, since it is based on a visual comparison between the theoretical curves and the FAD of the medium and large landslides. Additionally, the existed definition of mL introduces uncertainty due to the ambiguity in both the physical explanation of the rollover and its functional form. Here we focus on earthquake-induced landslides (EQIL) and aim to provide a rigorous method to estimate the mL and total landslide area of EQIL. We have gathered 36 EQIL inventories from around the globe. Using these inventories, we have evaluated existing explanations of the rollover and proposed an alternative explanation given the new data. Next, we propose a method to define the EQIL FAD curves, mL and to estimate the total landslide area. We utilize the total landslide areas obtained from inventories to compare them with our estimations and to validate our methodology. The results show that we calculate landslide magnitudes more accurately than previous methods.
Original languageEnglish
Number of pages1
Publication statusPublished - 2016
EventAGU Fall meeting 2016 - San Francisco, United States
Duration: 12 Dec 201616 Dec 2016


ConferenceAGU Fall meeting 2016
Country/TerritoryUnited States
CitySan Francisco
Internet address

Cite this