Landslide sediment transfer in formerly glaciated, mountain basins

We examine basin-wide colluvial sediment dynamics in the Tsitika and Eve Rivers(about 600 km2), coastal British Columbia, Canada. The colluvial sediment cascade is documented by classifying landslide types, characterizing the dominant sediment sources, and identifying preferential sites of colluvial delivery across the landscape. The study is based on the compilation and analysis of a 70-year landslide inventory. This analysis reveals that open-slope landslides delivering material to seasonal or perennial channels and fluvial terraces are the dominant source-to-sink pathways. This finding indicates high instability of the headwaters drainage network and its ongoing adjustment after generalized sediment recharge occurred in the Last Glacial Maximum (LGM). To quantify interactions between colluvial and fluvial processes the landscape has been subdivided into geomorphic domains, which include planar slopes, unchannelled valleys, erosional (source) and depositional (sink) colluvial channels, as well as fluvial channels. In the study period, landslide activity across landscape components has generated net sediment degradation on planar slopes and source colluvial channels, whereas unchannelled valleys, sink colluvial channels and fluvially-dominated channels have been accumulating material. The newly-developed scaling relation of the landslide sediment yield appears to be controlled by the spatial organization of relict glacial structures. It follows that landslide yield is highest in unchannelled topography, decreases at the scale of channel initiation(drainage area (Ad) about 0.002 km2), and remains constant for scales where source colluvial and hanging fluvial domains overlap (0.002 < Ad< 0.06). Landslide sediment injections start declining consistently beyond areas larger than 0.6 km2 (the scale of relict trough initiation), where fluvial environments are still partially connected to adjacent hillslopes. Cumulative daily yield indicates that landslide sediment redistribution across the landscape is limited to relatively small drainage areas;specifically, 90% of the colluvial load is released at scales smaller than about 0.6 km2. Evaluation of the studied landslide yield trend in conjunction with British Columbia fluvial sediment yield allows us to identify landscape scales of sediment aggradation and degradation, hence to infer preferential scales of contemporary sediment storage