The future of landslides' past-a framework for assessing consecutive landsliding systems

Landslides often happen where they have already occurred in the past. The potential of landslides to reduce or enhance conditions for further landsliding has long been recognized and has often been reported, but the mechanisms and spatial and temporal scales of these processes have previously received little specific attention. Despite a preponderance of qualitative and anecdotal evidence, analysis has been limited. As a result, there is little consensus on the meaning of terms such as landslide repetition, recurrence, and reactivation. This source of confusion is evident when such terms are also used to describe systems where landsliding is prevalent but unrelated to landslide history. Recent findings, partly based on a rare multi-temporal landslide inventory for an area in Italy, show that the impacts of earlier landslides affect a substantial fraction of landslides, that landslides following earlier landslides differ from those that do not, and that accounting for the effect of previous landslides can improve susceptibility assessments. These findings await confirmation in other landslide-prone landscapes but show that consecutive landsliding deserves more attention, which requires consistent terminology. No such terminology is presently available, and we therefore propose it in this manuscript. We use the term uncorrelated landsliding to describe situations where landslides are common, but where a correlation with environmental variables such as terrain steepness is not implied. We propose correlated landsliding to describe situations where landslides are common and correlations with environmental variables exist, and path-dependent landsliding to describe situations where causal relations exist between consecutive landslides, for instance, when landslides occur at the scarp of previous landslides. These are situations where past landslides impact future landslides. Within the path-dependent category, we distinguish three subcategories based on the spatial distance between earlier and later landslides: reactivation or continuation if essentially the same material recommences or continues to slide, local activation if an earlier slide causes changes in a local hillslope that cause a later slide, and remote activation if an earlier slide causes changes elsewhere in the landscape that cause a later landslide. We use this proposed set of terms to outline some prominent knowledge gaps and potential research questions.