Effects of terrain on temporal changes in susceptibility of debris flows and associated hydrogeomorphic processes after forest harvesting

Abstract

Forest harvesting is one of the most common and significant land cover changes and largely affects hydrogeomorphic processes in mountainous areas. Many studies have reported the direct impacts of forest harvesting on the susceptibility of slopes to landslides and surface erosion. However, ambiguity still remains in the impacts of forest harvesting on the occurrence of debris flows in channel. The impact of forest harvesting on the susceptibility of torrents to debris flows and associated hydrogeomorphic processes was interpreted in the Sanko catchment, central Japan, where the forest management record dates back to 1912. In order to clarify the contribution of landslides on the initiation of debris flows, landslides were classified into three types based on the termination point of the landslide sediment. Effects of terrain on the susceptibility of debris flows coming directly from landslides and those caused by mass movement of channel deposits were analyzed separately using aerial photographs taken in nine different years. Impact of forest harvesting on debris-flow occurrence was greatest in forest stands that were clearcut 1 to 10 yr earlier, with progressively lesser impacts continuing up to 25 yr after harvesting. The frequency of debris flows, which came directly from landslides, showed a similar time variation to that of landslides. A higher ratio of landslides turned directly into debris flows in steeper channel sections than in gentler channel sections. Angle of the landslide-channel junction also affected the ratio of landslides turning into debris flows. Debris flows caused by mass movement of channel deposits during high streamflow events also frequently occurred within 25 yr of clearcutting, possibly associated with active sediment supply from hillslopes by the landslide. Debris flows classified into this type were mostly initiated in channel sections steeper than 20°. Therefore, the effect of forest harvesting on debris flow susceptibility is variable depending on catchment terrain. Field monitoring showed that volume and spatial distribution of channel deposits, which affects bedload rate, depends on debris flow history in the catchment. Suspended sediment rate was also affected by debris flow and landslide history, because debris flows create sources of suspended sediment, such as bare areas along channel banks. Therefore, the impacts of debris flows on hydrogeomorphic processes after forest harvesting likely continue beyond the decline in debris flow and landslide frequency, due to changes to the sediment storage and sediment source by the occurrence of debris flows.