Wright, Charles M. 2023. Effects of Forest Harvest, Floods, and Wildfire on Bedload Export from Headwater Catchments in the H.J. Andrews Experimental Forest, 1957 – 2022. Corvallis: Oregon State University. 115 p. M.S. Thesis.
This study examined how annual bedload export volume and bedload characteristics were related to disturbances including logging, floods, debris slides, and wildfires over 48 to 65-year periods in small, steep catchments in conifer forests of the western Cascade Range, Oregon. Bedload – the material rolling, sliding, or saltating along the stream bed – is a key indicator of landscape erosion, but few studies document multi-decade patterns of bedload export and its relationship to landscape disturbance. This study took advantage of long-term measurements of bedload export from five headwater catchments in the steep, forested H.J. Andrews Experimental Forest. Bedload was measured annually in sediment basins in five monitored catchments ranging in size from 9 to 101 ha, over the period 1957 to present, and bedload samples were collected in 2020, 2021, and 2022. In the 1950s, all five catchments were dominated by mature and old-growth forest regenerated after wildfire in the early 1500s and mid 1800s. Two catchments (WS01 and WS10) were 100% clearcut in the 1960s and 1970s, and one catchment (WS03) was 25% patch clearcut in 1963 after roads were constructed in 1959. Major floods occurred in 1964/1965 and 1996. By 2020, clearcuts in WS01, WS03, and WS10 were 50- to 60-year-old conifer plantations. In September 2020, wildfire burned through WS01, WS02, and WS09.
Bedload export volume was related to maximum daily flow magnitude in unlogged catchments (WS02, WS09) in the period before wildfire (2020). Bedload export volume increased following clearcutting of mature and old-growth forest (WS01, WS10), patch cutting with roads (WS03) in the 1960s and 1970s, and after the 2020 wildfire (WS09). Surveys of stream channels of WS09 and WS10 in summer 2022 revealed that in-channel wood in WS09 (unlogged) stored nearly eight times more sediment than in WS10, where logging in 1975 and debris flows in 1986 and 1996 had removed large wood and depleted sediment storage in the channel. Average annual bedload export from WS09 (9 ha, burned) increased from 0.093 m3/ha before wildfire to 0.768 m3/ha after wildfire, while bedload export from WS10 (10 ha, unburned) was 0.031 m3/ha during these years. Physical characterization (texture, organic matter content, and char content) of bedload samples from WS09 (burned) and WS10 (unburned) indicate that following the wildfire disturbance, a number of factors including post-fire mortality, unloading of sediment stored in channels, and hillslope ravel processes increased organic and mineral material in bedload exported from burned catchments. Bedload export from small, steep, forested catchments continued to respond for more than half a century after human and natural disturbances. These findings illustrate general principles that can be extended to predict how catchments may respond to future forestry practices and other disturbances including wildfire.