Forecast of landslide inundation from precursory creep

Forecasting landslide dynamics upon catastrophic failure is crucial for reducing 10 casualties, yet it remains a long-standing challenge owing to their complex nature. Recent global studies indicate that catastrophic hillslope failures are commonly preceded by a period of precursory creep, motivating a novel scheme to foresee their hazard. Here, we introduce an approach to forecast landslide inundation by linking satellite-observed precursory displacements to models of consequent granular-fluid flows. We present its application to the 2021 Chunchi, 15 Ecuador landslide, which failed catastrophically and evolved into a mobile debris flow after four months of precursory creep, destroying 68 homes along its lengthy flow path. Underpinned by comprehensive uncertainty quantification and in-situ validations, we highlight the significance and feasibility of assessing landslide inundation hazard using precursory observables.