Control of Flow Sequence and Spatial Distribution of Debris Flow Input On River Network Modeling

Debris flows can have long-term effects on a watershed as deposited sediment at the upstream end of a river network can act as a sediment supply source for decades to centuries. Therefore, long-term simulation is critical to predict the combined effects of flow magnitude, duration, sequence, and intermittency for debris flow sediment routing. While such modeling scope in large spatial and temporal scale is often restricted by computational capacity, simplifying the flow hydrograph can help make the modeling tractable. Along with a 30-year daily flow simulation, this study explores the control of flow sequence and constant flow on debris flow transport through Provo River network at different time horizons. We focused on the Provo River Watershed upstream of the Jordanelle Reservoir in Utah. Additionally, this work also investigates the effect of differences in spatial distribution of debris flow sediment input to the network by analyzing corresponding tributary and mainstem characteristics. Simulation results from the reduced hydrographs (of constant flow and different sequences) show that these can produce long-term transport comparable to the original flow record. Although the initial (1-5 years) discrepancy is high, these differences decrease over time (after 10 years). The effect of flow sequences was less important for sand than gravel, as both high and low flow would move sand initially. This eventually results in high total transport when the low flow years are followed by high flow years, because later high flows can move the coarse gravel after the early sand removal by low flows. Because the simple compressed hydrograph approximated long-term transport, we employed a constant flow hydrograph to investigate the network characteristic controls on debris flow sediment transport. Model results showed the importance of drainage area ratio between the tributary and mainstem on storage (in mainstem and reservoir), similar to previous studies. When mainstem slope was similar to the tributary slope, the resulting overall transport was more efficient. This study also suggests that the extent of mainstem aggradation depends mainly on mainstem slope properties. Such a network-scale modeling study quantitatively identifies geomorphic significant tributaries, which are important for river biodiversity. Besides, this study focuses on how the results from a reduced hydrograph vary from long-term records at different timescales. With the expected future increase of magnitude and frequency of high floods, and the potential of increased severity and frequency of extreme events due to climate change, the long-term simulation of flow sequences can inform river managers about how to better prepare to reduce loss from debris flows, and to improve overall river and watershed management.