Transient changes of landslide rates after earthquakes

Abstract

Earthquakes impart an impressive force on epicentral landscapes, with immediate catastrophic hillslope response. However, their legacy on geomorphic process rates remains poorly constrained. We have determined the evolution of landslide rates in the epicentral areas of four intermediate to large earthquakes (M w , 6.6-7.6). In each area, landsliding correlates with the cumulative precipitation during a given interval. Normalizing for this meteorological forcing, landslide rates have been found to peak after an earthquake and decay to background values in 1-4 yr, with the decay time scale probably proportional to the earthquake magnitude. The transient pulse of landsliding is not related to external forcing such as rainfall or aftershocks, and we tentatively attribute it to the reduction and subsequent recovery of ground strength. Observed geomorphic trends are not linked with groundwater level changes or root system damage, both of which could affect substrate strength. We propose that they are caused by reversible damage of rock mass and/or loosening of regolith. Qualitative accounts of ground cracking due to strong ground motion abound, and our observations are circumstantial evidence of its potential importance in setting landscape sensitivity to meteorological forcing after large earthquakes.

Do landslides follow landslides? Insights in path dependency from a multi-temporal landslide inventory

“…The decreasing temporal effect of earlier landslides on susceptibility is to some extent comparable with the legacy effect of earthquakes on the occurrence of earthquake-induced landslides (Lin et al 2007;Marc et al 2015;Parker et al 2015). For instance, Marc et al (2015) demonstrated that landslide rates were significantly elevated within 0.7-4.5 years after four earthquakes with magnitude higher than 6 Richter.…”

The control of earthquake sequences on hillslope stability

“…Our results are most applicable to hillslope deposits of similar rheology and stress history to those tested here, indicative of shallow (~10 to 20 m depth) failures in near-surface slope materials, within which prefailure strain, manifest as ground cracking, is frequently evident. Our findings may also be applicable at the broader "landscape scale", where seismically controlled dilation controls the frictional strength of heavily damaged rock masses [Marc et al, 2015;Scheingross et al, 2013]. We have demonstrated that lower magnitude ground-shaking events can in some cases cause progressive densification of sediment, increasing frictional strength and reducing susceptibility to landsliding during subsequent seismicity and also in response to precipitation events.…”

“…As such, our understanding of both coseismic and interseismic landsliding rates and patterns is primarily empirical which potentially limits the transferability of patterns and transient rates of landsliding identified. Marc et al [2015] reported an exponential reduction in elevated landslide activity and a return to preseismic landslide rates over subdecadal timescales. The four case studies considered by Marc et al [2015] share broadly similar seismogenic settings, and all experienced an earthquake sequence characterized by a mainshock and a series of aftershocks.…”

“…Marc et al [2015] reported an exponential reduction in elevated landslide activity and a return to preseismic landslide rates over subdecadal timescales. The four case studies considered by Marc et al [2015] share broadly similar seismogenic settings, and all experienced an earthquake sequence characterized by a mainshock and a series of aftershocks. While it is effectively impossible to know the stress history of any slope, our results suggest that local fault characteristics will ultimately generate characteristic slope responses at a landscape scale evident over decades to centuries.…”

“…In addition to causing extensive socioeconomic disruption [Marano et al, 2010], earthquake-induced landslides play a key role in the evolution of mountain landscapes, increasing sediment flux through the fluvial network [Dadson et al, 2004;Hovius et al, 2011] and contributing to net erosion rates [Marc et al, 2015;Parker et al, 2011;Robinson et al, 2016]. While earthquake ground shaking triggers near-instantaneous landsliding [Li et al, 2014], some slopes do not fully fail and are weakened [Khattak et al, 2010], resulting in elevated susceptibility of hillslopes to landsliding during postseismic rainfall [Lin et al, 2008] and subsequent seismicity [Parker et al, 2015].…”