10:45 AM - 12:15 PM
[MIS07-P01] Experimental investigation of mudflow depositional processes
Keywords:Debris flow, Mudflow, Flume experiment, Depositional processes, Two-layer model
Depositional processes determine the inundation area of a debris flow. In coarse-grained debris flows (stony debris flows), sediment deposition occurs as the current bed slope approaches an equilibrium bed slope corresponding to the current sediment concentration. Compared with stony debris flows, fine-grained debris flows (mudflows) tend to have to gentler slopes. This difference may be related to particles that are supported by turbulence in mudflows; however, it cannot be fully explained by the theory of low-concentration turbulent suspension flows. Therefore, this study examines the mechanism of mudflow deposition using flume experiments.
A straight rectangular channel 8 m in length and 0.1 m in width was used for the experiment. Water and sediment were mixed by agitation in a tank located downstream of the channel, and the mudflow was input at the upstream end using a sand pump. The supplied mud flows down the channel and returns to the tank from the downstream end in a circulating system, allowing steady flow at a constant flow rate and sediment concentration in the absence of deposition. In each experiment, the flow rate and sediment concentration of the mudflow were varied experimentally; the channel gradient of sediment deposition and its formation processes were examined. The initial slope of the channel was sufficiently large to prevent sediment deposition, and the slope was reduced in increments of 0.1° until deposition occurred. The sediment particles used in the mudflow were Tohoku silica sand No. 8, with a median grain size of 0.11 mm.
The following processes of mudflow deposition were observed in the experiment. From the initial gradient to the gradient at which sediment deposition occurred, mud flowed over the rigid bed without forming a deposition layer. However, as the channel gradient was lowered to a specific level, a deposition layer began to form at the downstream end of the channel. Then, when the channel gradient was kept constant, a deposition layer gradually developed in the upstream direction. Subsequently, because a hydraulic jump occurred where the bed gradient was locally lower, sediment deposition progressed upstream in the subcritical flow portion of the deposition layer; conversely, erosion occurred in the downstream supercritical flow portion, resulting in upstream migration of the deposition layer. This deposition layer eventually reached the upstream end and the process repeated, such that it recurred downstream. We assumed that this process represented sand waves; thus, we assessed whether the channel gradient at the beginning of deposition could be considered equivalent to the equilibrium bed slope of a stony debris flow.
Comparison of the sediment concentration in the mudflow and the channel gradient at the beginning of deposition revealed an overall trend whereby a greater sediment concentration caused a greater channel gradient at the beginning of deposition. However, in contrast to stony debris flows (where one-to-one correspondence exists between sediment concentration and equilibrium bed slope), we found that at similar sediment concentrations, greater flow depth was associated with a lower channel gradient at the beginning of deposition. Next, the theoretical equilibrium mudflow bed gradient was compared with values measured in this experiment. Similar to the derivation of an equilibrium bed slope for stony debris flows, the slope for mudflows was calculated considering the balance between external forces and shear stresses at the bed. A two-layer model containing a boundary within the flow at a specific height above the bed was used. Above this boundary, sediment particles were suspended because of turbulence; below it, particles flowed in a laminar state, forming a layer in which inter-particle stresses were dominant. The channel gradient at the beginning of deposition generally corresponded well with the theoretical equilibrium mudflow bed slope. Based on these findings, the equilibrium bed slope obtained from a two-layer model can be used to estimate the gradient at which mudflow deposition occurs.
A straight rectangular channel 8 m in length and 0.1 m in width was used for the experiment. Water and sediment were mixed by agitation in a tank located downstream of the channel, and the mudflow was input at the upstream end using a sand pump. The supplied mud flows down the channel and returns to the tank from the downstream end in a circulating system, allowing steady flow at a constant flow rate and sediment concentration in the absence of deposition. In each experiment, the flow rate and sediment concentration of the mudflow were varied experimentally; the channel gradient of sediment deposition and its formation processes were examined. The initial slope of the channel was sufficiently large to prevent sediment deposition, and the slope was reduced in increments of 0.1° until deposition occurred. The sediment particles used in the mudflow were Tohoku silica sand No. 8, with a median grain size of 0.11 mm.
The following processes of mudflow deposition were observed in the experiment. From the initial gradient to the gradient at which sediment deposition occurred, mud flowed over the rigid bed without forming a deposition layer. However, as the channel gradient was lowered to a specific level, a deposition layer began to form at the downstream end of the channel. Then, when the channel gradient was kept constant, a deposition layer gradually developed in the upstream direction. Subsequently, because a hydraulic jump occurred where the bed gradient was locally lower, sediment deposition progressed upstream in the subcritical flow portion of the deposition layer; conversely, erosion occurred in the downstream supercritical flow portion, resulting in upstream migration of the deposition layer. This deposition layer eventually reached the upstream end and the process repeated, such that it recurred downstream. We assumed that this process represented sand waves; thus, we assessed whether the channel gradient at the beginning of deposition could be considered equivalent to the equilibrium bed slope of a stony debris flow.
Comparison of the sediment concentration in the mudflow and the channel gradient at the beginning of deposition revealed an overall trend whereby a greater sediment concentration caused a greater channel gradient at the beginning of deposition. However, in contrast to stony debris flows (where one-to-one correspondence exists between sediment concentration and equilibrium bed slope), we found that at similar sediment concentrations, greater flow depth was associated with a lower channel gradient at the beginning of deposition. Next, the theoretical equilibrium mudflow bed gradient was compared with values measured in this experiment. Similar to the derivation of an equilibrium bed slope for stony debris flows, the slope for mudflows was calculated considering the balance between external forces and shear stresses at the bed. A two-layer model containing a boundary within the flow at a specific height above the bed was used. Above this boundary, sediment particles were suspended because of turbulence; below it, particles flowed in a laminar state, forming a layer in which inter-particle stresses were dominant. The channel gradient at the beginning of deposition generally corresponded well with the theoretical equilibrium mudflow bed slope. Based on these findings, the equilibrium bed slope obtained from a two-layer model can be used to estimate the gradient at which mudflow deposition occurs.