Influence of connectivity configurations and moisture on landslide mobility: 2018 Eastern Iburi earthquake Hokkaido, Japan

Fatalities associated with landslides triggered by rainfall (LR), or earthquakes (LE) are strongly related to landslide mobility. Many studies have shown that the extent of landslide mobility is related to extrinsic factors such as connectivity, which is defined as connected deposits of multiple landslide sources or scars (e.g., coalesced landslides). Meanwhile, intrinsic factors such as high-water availability on concave hillslopes can affect mobility, considering that water in the soil initiates soil collapse and transport downslope. However, research on these extrinsic and intrinsic factors that affect landslide mobility have only been conducted for LR. Interestingly, we found that landslide mobility triggered by the 2018 East Iburi earthquake manifested in concave topography, more typical for LR, which may be related to different types of connectivity of sediment in the channel. Therefore, the objectives of this study were to characterize landslide mobility in LE cases based on the types of connectivity and to examine the possible impacts of topographically controlled soil moisture on landslide mobility. To examine the landslide mobility, we evaluated the horizontal runout distance (L), the ratio between the highest landslide initiation point (H) and L (H/L ratio), and the ratio of the landslide scar area (S) to the total landslide area from the scar initiation point to the toe (F) (S/F ratio). When landslides from multiple sources connect into the channel and form a single deposit, S will be defined as the cumulative scar area whereas H and L will be based on the highest scar point to the end of the lowest toe point within entire delineated F area. We assessed landslide mobility within a 58.6 km2 area in the middle of the Atsuma watershed, Hokkaido. Using post-LiDAR shaded relief (0.5 m), we identified 203 landslide scars (S) within 85 composite landslide areas (F). Among those 85 samples, 61% of them mobilized <250 m. Nonetheless, Iburi landslides had small H/L and S/F ratio values of <0.5 (85% and 45%, respectively), indicating highly mobilized landslides. To correlate with connectivity, we classified landslides into six different types including: (1) single-scar; (2) coalesced-scars (connected deposits from different sources in different catchments); (3) amalgamation-scars (connected deposits from adjoining sources); (4) headwater-scars (sources from headwater hillslopes only); and (5) ladder-scars (connected deposits from vertically adjacent sources), and (6) complex-scars (combination of two or more types mentioned above). As expected, the complex-scars landslides travelled the greatest distances (mean L: 667 m; mean H/L: 0.23; mean S/F: 0.48), because the higher number of scars and larger volume of cumulative sediment increased runout length. Interestingly, we also found several single-scar landslides type may also trigger longer mobility (>500 m) than combined landslides, indicating a possible effect of the presence of water on concave slopes. Future research will focus on compiling a more complete landslide inventory including S and F emphasizing types of connectivity. In addition, topographically controlled soil moisture analysis will be examined to assess possible effect of water in hillslopes.