5:15 PM - 6:30 PM
[HCG28-P03] Flume experiment on crevasse-splay deposits formed by the 2019 flooding of the Chikuma River
On October 13, 2019, flooding of the Chikuma River was caused by heavy rain in central Japan. The 30 m wide levee revetment of the left bank of the river was breached over a length of 70 m in Hoyasu, Nagano City. As a result, gravelly and sandy piles were formed at upstream and downstream sides of the crevasse channel perpendicular to the river. This study aims to obtain a better understanding of the processes involved in transport and deposition during the flooding by flume experiments to reproduce the distribution and sedimentary characteristics of the surveyed crevasse-splay deposits.
In the flume experiments, we first mixed two types of sand grains with water, and flow it into the acrylic experimental flume which was a simplified scale model of the Chikuma River (about 2 m long and 8 cm wide). We used green-colored sand (150–250 μm in diameter) and red-colored sand (40–100 μm in diameter) as analogs of gravels and sand grains in field scale respectively. The flow depth in the main channel was controlled by narrowing the flume width at the downstream end with the board. After the flow reached the steady condition, we removed 2 cm high boards piled to compose the sluice gate (35 cm) at the middle part of the water way, causing levee breach on the floodplain (100×150 cm) adjacent to the water way. We conducted ten runs of the experiment with different conditions. Sediment sampling and analysis were conducted on Run 10 because of its highest reproducibility of the actual levee breach.
On Run 10, we first removed the top of four small boards to cause overspilling of the flow at the time of 59 s after the start of experiment. At the time of 28 s after overspilling, we removed two boards to cause the successive levee breaching while the last board was left until the time of 132 s after levee breach. Based on the video observation, a crevasse channel was formed the side sluice gate with deposition of piles composed of red sands on both sides of the channel during overflow stage. The direction of the crevasse channel was inclined downstream by 24° measured from the direction perpendicular to the main channel. Immediately after levee breach, the flooding flow swept away the sediment initially deposited on the bottom, and a pile of sediment was deposited instantly in the upstream region of the crevasse channel. As a result, the channel shifted its direction toward downstream (32° at 30 s after levee breach) because the sediment pile interrupted the flow. At the time of 90 s after the onset of levee breach, the depositional area in the upstream region of the channel expanded, and the channel consequently shifted to downstream direction furthermore (38°).
The measurement of the flooding flow indicated that the flow velocity immediately after levee breach was the highest during the experimental period. The flow in the overbank area was about 41 cm/s before levee breach and was about 83 cm/s at the timing when levee breach occurred. Then, it became about 53 cm/s at 60 s after levee breach. The pile deposited in the upstream side of the crevasse channel was formed from the bedload of the flooding flow in the initial stage of levee breach, and thus it presumably exhibited high bed shear stress. Subsequently, the pile was isolated at the upstream side by the shift of the crevasse channel to downstream region with waning of the flow velocity.
The feature of the coarse-grained pile in the upstream side of the crevasse channel quite resemble that of the gravel pile formed in the 2019 Chikuma River flooding. Thus, it can be interpreted that the crevasse splay deposits at the study site were formed in the similar process in which the shift of crevasse channel occurred with waning the flow. Further experimental and field surveys are needed to reveal whether the feature of the coarse-grained sediment pile examined in this study is a general characteristics of the crevasse splay deposits or it can be only formed at a specific hydraulic condition.
In the flume experiments, we first mixed two types of sand grains with water, and flow it into the acrylic experimental flume which was a simplified scale model of the Chikuma River (about 2 m long and 8 cm wide). We used green-colored sand (150–250 μm in diameter) and red-colored sand (40–100 μm in diameter) as analogs of gravels and sand grains in field scale respectively. The flow depth in the main channel was controlled by narrowing the flume width at the downstream end with the board. After the flow reached the steady condition, we removed 2 cm high boards piled to compose the sluice gate (35 cm) at the middle part of the water way, causing levee breach on the floodplain (100×150 cm) adjacent to the water way. We conducted ten runs of the experiment with different conditions. Sediment sampling and analysis were conducted on Run 10 because of its highest reproducibility of the actual levee breach.
On Run 10, we first removed the top of four small boards to cause overspilling of the flow at the time of 59 s after the start of experiment. At the time of 28 s after overspilling, we removed two boards to cause the successive levee breaching while the last board was left until the time of 132 s after levee breach. Based on the video observation, a crevasse channel was formed the side sluice gate with deposition of piles composed of red sands on both sides of the channel during overflow stage. The direction of the crevasse channel was inclined downstream by 24° measured from the direction perpendicular to the main channel. Immediately after levee breach, the flooding flow swept away the sediment initially deposited on the bottom, and a pile of sediment was deposited instantly in the upstream region of the crevasse channel. As a result, the channel shifted its direction toward downstream (32° at 30 s after levee breach) because the sediment pile interrupted the flow. At the time of 90 s after the onset of levee breach, the depositional area in the upstream region of the channel expanded, and the channel consequently shifted to downstream direction furthermore (38°).
The measurement of the flooding flow indicated that the flow velocity immediately after levee breach was the highest during the experimental period. The flow in the overbank area was about 41 cm/s before levee breach and was about 83 cm/s at the timing when levee breach occurred. Then, it became about 53 cm/s at 60 s after levee breach. The pile deposited in the upstream side of the crevasse channel was formed from the bedload of the flooding flow in the initial stage of levee breach, and thus it presumably exhibited high bed shear stress. Subsequently, the pile was isolated at the upstream side by the shift of the crevasse channel to downstream region with waning of the flow velocity.
The feature of the coarse-grained pile in the upstream side of the crevasse channel quite resemble that of the gravel pile formed in the 2019 Chikuma River flooding. Thus, it can be interpreted that the crevasse splay deposits at the study site were formed in the similar process in which the shift of crevasse channel occurred with waning the flow. Further experimental and field surveys are needed to reveal whether the feature of the coarse-grained sediment pile examined in this study is a general characteristics of the crevasse splay deposits or it can be only formed at a specific hydraulic condition.