Suspended Sediment Loading from Timber Harvest and Landslides, Railroad Gulch Paired Watershed Experiment, Elk River, California, USA

A watershed experiment evaluated effectiveness of best management practices for avoiding sedimentation from timber harvest from 2014-2020. The site was Railroad Gulch, a 2.7 km2 sub basin of the Elk River, Northern California. Continuous data (water depth and turbidity) pump sampling and velocity measurements were used to quantify sediment loading. The treatment reach was selectively harvested in 2016, using cable yarding, after road reopening and construction in 2015. The control branch was undisturbed; however, a significant landslide occurred in 2017. Sediment yields on the treatment branch ranged from 49 t km2 during the 2014 drought year to 1178 t km2 during 2017, the third wettest year since 1888. Sediment yields in the control branch also increased from dry to wet years, with 37 t km2 in 2014, and 774 t km2 in 2017. The largest sediment yield (950 t km2, 2019), was in response to landslide-derived inputs. The long-term erosion estimate using Be-10 analysis was 302 t km-2 yr-1. While sediment loading from harvest was evident, some of the increase may have resulted from high precipitation and greater treatment branch drainage density. Sediment loading from timber harvest and landsliding recovered quickly to pre-disturbance levels. Changes in river cross sectional area, bank erosion and bed material indicated scouring and coarsening trends, with scour during wetter years and trends in drier years. A clear signal was evident from landslides but not timber harvest. Road surveys indicated increases in rill erosion and plume deposition on native surfaces after roadwork, that stabilized by revegetation following timber harvest. Sediment fences gave road erosion estimates of 0.0 kg m2 yr-1 to 4.8 kg m2 yr-1. Roads with rilled surfaces had two-fold higher depositional plume lengths compared to roads without. Roads with native surfaces experienced greater surface erosion than those with rock-surfaces. Road crossings resulted in consistent increases to stormflow turbidity in watercourses at several locations. Restoration of crossings appeared to reduce impacts at several locations. Monitoring indicated that the channel network did not expand from extremely wet years nor timber harvest. No road or timber harvest-related landslides occurred during the study. Peak discharge did not appear to increase from timber harvest.