Patterns and processes of sediment transport following sediment-filled dam removal in gravel bed rivers

Graduation date: 2006Dam removal is increasingly viewed as a river restoration tool because dams affect so many aspects of river hydrology, geomorphology, and ecology; but removal also has impacts. When a dam is removed, sediment accumulated over a dam’s lifetime may be transported downstream; and the timing, fate and consequences of this sediment remain some of the greatest unknowns associated with dam removal. In this thesis, I develop a conceptual model for erosion and deposition following removal of sediment-filled dams in mountain streams, and use field studies to document actual change. The data show that reservoir erosion in mountain rivers is likely to occur by knickpoint migration, with 85% of stored sediment being released during a single storm event in two field studies, at shear stresses less than that required for mobilization of the median surface particle size. Coarse sediment is predicted to deposit close to the dam with channel aggradation decreasing exponentially with increasing distance downstream, although some channel features are shown to have a greater propensity for aggradation than others. Field studies show that turbidity associated with dam removal and reservoir erosion may decrease hyporheic exchange, but gravel deposition (e.g., 470 m3 of gravel from Dinner Creek Dam) has the potential to more than offset that decrease, and increased hyporheic exchange is shown to reduce diurnal temperature change. Macroinvertebrate density and taxa richness did not respond to dam removal itself, but rather with time-lagged reservoir erosion. Following reservoir erosion, macroinvertebrate density recovered quickly, although longterm taxa community composition appears to be altered. On the Sandy River, field measurements of shear stress and patterns of sediment deposition following cold lahars were used as an analog to predict the fate of fine sediment, which is likely to deposit far from the dam. Results show that the Sandy River has little capacity for fine sediment storage in pools above RK 6.4 (~ 42 kilometers below Marmot Dam) at discharges associated with reservoir sediment releases. Taken as a whole, this paper illustrates a complex suite of process that may accompany removal of sediment-filled dams in mountain rivers