Capturing the legacy effect of the ground motion in the spatial distribution of rainfall-induced landslides

Abstract:

Earthquakes elevates landslide susceptibility in the post-seismic period and play a major role in rainfall-triggered landslide events. However, the coupled effect of earthquake and rainfall is rarely investigated in landslide susceptibility assessments although it may be crucial to predict landslide occurrences over space and time. Here, we investigate this concept by accounting for the earthquake’s legacy in rainfall-induced landslide (RFIL) scenarios. We hypothesize that the Peak Ground Acceleration map of the most recent and the largest earthquake, together with rainfall proxies associated with RFIL can be simultaneously fitted in a landslide susceptibility model to capture their coupled effect. To test this hypothesis, we used a Generalized Linear Model underlain by a Bernoulli probability distribution and examined two time series of landslides, including earthquake- and rainfall- induced landslide inventories mapped in Indonesia. First, we present how the contribution of earthquake legacy effects decays through time from co-seismic to post-seismic periods. Second, we show how the legacy of a previous earthquake can still control the spatial distribution of RFIL. Specifically, in the examined cases, we observe that the signal of ground shaking is quite distinct immediately after an earthquake and therefore the first few extreme rainfall events following the earthquake are the most likely cases where the legacy effect plays a role in RFIL distribution. Our findings also show that based on remote observations, the signal appears to disappear within approximately three years.