Rainfall-induced shallow landslides represent a ubiquitous natural hazard in mountainous regions (Sidle & Ochiai, 2006) and exert critical control on sediment transfer, channel morphology, and landscape evolution (Hovius et al., 2000;Korup et al., 2007;Larsen et al., 2010). Evidence suggests that landslides are more likely to occur in historical landslide areas with favorable local climate and topography where soil depth is modified by landslide stripping and subsequently restored by soil regeneration (Mirus et al., 2017;Samia et al., 2017). Rainfall-induced shallow landslides typically strip the soil mantle and expose the underlying bedrock. The reduced soil depths in the stripped region may recover with soil regeneration rates that are primarily determined by soil development (through pedogenesis and bedrock weathering, often termed "soil production") and soil transport which are affected by local climate and biological perturbation (e.g., Dietrich et al., 1995;Heimsath et al., 1997). The transport of debris from the surrounding hillslopes to bedrock hollows and to channels is caused by different mechanisms such as soil creep, animal burrowing, and tree throw. Even though soils do not physically diffuse due to molecular interactions such as with heat or solutes, the slow local spreading of soil material resembles diffusion in landscape scale transport models expressed as soil diffusivity (Reneau & Dietrich, 1991), hence soil diffusivity is not a soil physical property, but a spreading parameter (e.g., Heimsath et al., 1997) reflecting soil properties such as grain size distribution and local factors such as bioturbation, climate and vegetation. The soil production by bedrock weathering is an outcome of complex interplay of physical, chemical, and biogenic process (Birkeland, 1999). The imprint of historical landslide stripping and subsequent soil regeneration on local topography and soil depth, in turn, influences the characteristics (e.g., timing, location, volume) of future landslide triggering and sediment production and delivery to fluvial systems (Crozier & Preston, 1999;Parker et al., 2016). In this paper, we term effects of past landslide stripping and subsequent soil regeneration on future landslide characteristics and related geophysical processes (e.g., fluvial sediment dynamics) as "landslide legacy effects" with cumulative historical landslides termed "legacy landslides." This resembles the notion of "path-dependent landsliding" proposed by Temme et al. (2020) for causal relations between consecutive landslides. However, "landslide legacy" combines imprints of past landslide events with subsequent soil regeneration over soil generation and depth evolution time scales. Hence, regions affected by "legacy landslides" are areas that experienced repeated mass wasting and subsequent soil regeneration in the past and are not limited to the neighborhoods of individual landslides triggered