Dynamic earth system and ecological controls of rainfall-initiated landslides

Rainfall-initiated landslides continue to inflict damages and loss of life throughout the world. Processes and mechanisms revealed from hydrological, geomorphic, geotechnical, pedological, geological, hydrochemical, and biological investigations have advanced our understanding of these effects on slope stability; however, the interactions amongst these processes and attributes as they affect the initiation and propagation of landslides are not aswell understood. Too often landslide studies are conducted from only one, or atmost, two of these perspectives.Moreover, while precipitation and hydrology are recognized as dynamic influences, the earth system and ecological effects are often assumed to be static.Weassess the interplay of these processes related to landslides triggered by positive pore water accretion and loss of soil suction. Each of these conditions arguably requires a different viewon the processes that cause slope failure and predictive approaches. This review starts from the perspective of dynamic, adapting earth and ecological systems and discusses how these attributes relate to landslide initiation, mode, location, and timing. Specifically, the role that large- and small-scale preferential flow plays in both contributing to and mitigating instability is elucidated, including effects of bedrock exfiltration.We also examine how and under what conditions these pathways manifest in different soils, lithology, and landforms. The multiple effects of rhizosphere processes on slope stability are discussed, including root reinforcement, evaporation fromcanopies and litter layers, transpiration, and the role of root structure affecting preferential flow paths. Rainfall-initiated landslides involve highly dynamic hydrologic, earth surface, and ecological processes that persist over a range of spatial and temporal scales; however, guidance for overcoming these challenges has been elusive. A conceptual framework is presented to shed light on these dynamic and interactive processes that should lend insights into why and when certain slopes fail during storms, while other seemingly similar slopes do not fail. Such advances will benefit landslide hazard assessments and disaster responsiveness protocols.