Sediment motion and velocity in a glacier-fed stream

Abstract

Current understanding of coarse sediment transport (e.g. threshold for motion, travel length and virtual velocity) in mountain rivers is still quite limited, and even less is known about glacial streams. However, the hydrological characteristics of these systems (strong daily discharge fluctuations, high water turbidity) pose challenges to the use of tracers to monitor bed sediment dynamics, as tagged clasts are usually located after bedload events when flow stage has receded, e.g. by means of portable antennas in the case of Passive Integrated Transponders (PIT). The use of stationary antennas, still scarcely in use worldwide, to detect PIT-tagged particles has potential advantages in glacier-fed streams. If water discharge is monitored continuously, a stationary antenna provides real time data on the actual discharge at the moment of tracer particles passage. This study focuses on incipient motion and virtual velocity of bed particles measured by a stationary antennas system in a steep mountain channel (Saldur River, drainage area 18.6km2, Italian Alps) where significant daily discharge fluctuations and bedload transport occur as a result of a nivo-glacial regime. Four stationary antennas were installed 50-m apart in the study reach. A total of 629 PIT-tagged clasts were inserted in the studied reach between 2011 and 2014, ranging in size from 35mm to 580mm, with an overall recovery rate of around 44%. Critical discharge for sediment entrainment was obtained by detecting the movement of tracers placed immediately upstream of antennas. Virtual velocity was derived by knowing distances between the antennas and travel time of tracers. Results on initiation of motion show that the relationship between the size of transported tracers and the discharge measured at the time clasts were passing the stationary antenna is very weak. The influence of antecedent flows on incipient motion was thus investigated by dividing the highest discharge recorded between each PIT deployment and the subsequent entrainment by the actual critical discharge at the time of movement (ratio Qmax/Qc). Results show that approximately 50% of tracers moved at Qmax/Qc≤1.2, and that 73% of tracers moved at Qmax/Qc<1.5. Therefore, about 30% of tracers had to previously experience a discharge substantially greater than the one that actually mobilized them. Also, coarser particles moved at higher Qmax/Qc ratios, suggesting that higher antecedent flows may be needed to destabilize bed clustering. Results on the virtual velocity of the PIT-tagged clasts employed in the field show that the virtual velocity turned out to be highly variable (ranging from 101 to 10−5mmin−1) and weakly related to either particle size or flow discharge. However, virtual velocity was well correlated with the highest flow discharge experienced by each tracer normalized by a percentile of the flow duration curve. This evidence further stresses the importance of flow history on sediment entrainment and transport. Finally, the pros and cons of the deployed monitoring technology are discussed.