EP038-0007
Quantifying Spatiotemporal Patterns of Grain Size, Roughness, and Flow Resistance in the Liwu River, Taiwan
Quantifying Spatiotemporal Patterns of Grain Size, Roughness, and Flow Resistance in the Liwu River, Taiwan
Friday, 11 December 2020
Poster
Abstract:
The incision of bedrock rivers into uplifting mountain ranges depends on the size and flux of sediment delivered from hillslopes, which can vary dramatically in time and space. The variation in sediment size and cover also leads to patterns in streambed roughness, which influences sediment transport and incision through flow resistance. Measuring bed roughness directly in mountain rivers is challenging and often assumed uniform in models even for locations with high variability in sediment supply. Here, we use a suite of Unmanned Aerial Vehicle (UAV) structure-from-motion photogrammetry surveys to characterize how streambed roughness varies spatially across over 10 kilometers of river channel and temporally over 5 years of repeat surveys in the Liwu River, Taiwan. For 3 reaches, we mapped grain size patterns using cm-scale orthoimagery and used these maps to calibrate interpretation of streambed roughness at multiple scales across the study area. Channel morphology ranges from bare bedrock exposure through cobbles and gravel bars to reaches with boulders larger than 10 m in diameter. Our results show that the standard deviation of roughness heights varies with median grain size despite having variations in sediment sorting. We see that roughness varies along the surveyed reaches by three orders of magnitude, with the highest roughness patches reflecting coarse sediment input from hillslopes. We also observe that roughness can change by two orders of magnitude in the same location over time, where gravel and cobbles fill in and reduce net protrusion of boulders delivered or exhumed by typhoon flows. Our results highlight the scales and patterns of grain size, roughness, and mobility at the reach scale, improving estimates of flow resistance in bedrock channels.