Armor development and bedload transport processes during snowmelt and flash floods using laboratory experiments, numerical modeling, and field-based motion-sensor tracers

Bedload transport rates are notoriously difficult to predict during floods, yet accurate predictions are necessary for assessment and mitigation of potential hazards, such as river avulsions and overall channel stability. The presence or absence of coarse bed surface layers, or armor, drastically affect bedload transport rates. Armor-forming and bedload transport processes are reasonably well understood in perennially-flowing gravel-bed rivers during base flows. In contrast, arguably less is known about how floods transport bed sediment. In particular, dryland, ephemerally-flowing rivers are often relatively unarmored compared to their perennial counterparts, yet the exact processes that prevent armor development is still an area of active research. Therefore, I used a combination of laboratory experiments, numerical modeling, and field work, to better understand bedload transport and armor development in flood flows. Laboratory experiments explored the role that high-velocity flood bores may play in the maintaining relatively unarmored bed surfaces typical of dryland rivers. Numerical modeling was then used to explore the effect hydraulic and sediment supply variables may have on the formation of armor. Finally, I used motion sensor tracers in a snowmelt-fed gravel-bed river to quantify changing bedload transport throughout a series of above-threshold diurnal flood events. I found flash flood bores likely do not help maintain relatively unarmored bed surfaces in dryland, ephemerally-flowing rivers. Instead, dryland rivers likely remain unarmored by both high sediment supply to the channel, and flows which cannot selectively transport and remove finer grains. Motion sensor tracer data show that the thresholds of particle motion, and thus bedload transport rates, change with history of above-threshold flow.