9:00 AM - 9:15 AM
[ACG43-01] Identification of the Factors Contributing to the Occurrence of Sandbars Based on Observational Big Data
★Invited Papers
Keywords:alternate bars, sediment transport, observation big data
In rivers composed of gravel and sand materials, periodic undulating structures known as sand waves spontaneously form on the riverbed. These are classified into three categories based on the scale of their shapes, with the largest of these being classified as sandbars. Numerous studies have been conducted to physically investigate sandbars, research using stability analysis has revealed that sandbars occur when the river width to depth ratio exceeds a certain value. In this analysis, small disturbances are introduced to the initial bed surface, and in this field, the cause of sandbar formation has been considered to be these small disturbances. However, no data has been presented to demonstrate that small disturbances are the cause of sandbar formation, and the mechanism of sandbar occurrence remains unclear. Moreover, methods for finely measuring the process of sandbar formation and development have not yet been established.
Numerical simulation has been used as the research method in recent studies of properties of sandbars. In the numerical simulation of model for flow and sediment transport, small disturbances are assumed to be a factor in the occurrence of sandbars, and small disturbances are given as initial conditions, as in stability analysis. Therefore, it is difficult to consider factors other than small disturbances in the occurrence of sandbars. In response to this problem, the authors developed a measurement system called Stream Tomography (ST), which can measure the geometrical shape of the water surface and bottom in a model experiment of the process of sandbar initiation and development with high spatial resolution and high temporal frequency, enabling the acquisition of observation big data equivalent or superior to numerical analysis. We first conducted a model experiment in which a sandbar was spontaneously occurred in a straight channel by initially spreading fine sand on a flat surface, measured the geometry of the water surface and bottom surface during the process of sandbar generation and development using ST, and analyzed the wave numbers in the flow direction and transverse direction for the water surface and bottom surface geometry. The results showed that the bottom surface remained flat immediately after the start of the model experiment, whereas the water surface had a wave train with a wave number of about 20. Next, we measured planar distribution of flow velocity by PTV using the aforementioned straight channel as a fixed bed. The results showed that the planar distribution of the velocity was equivalent to that of the water surface at the initial time of the model experiment with the moving bed, and that standing waves existed. Generally, it is said that in a straight and flat channel, the flow velocity and flow depth are spatially uniform and isentropic, but the detailed measurements revealed that standing waves exist even in a flat bed. These results suggest that one of the factors occurring sand waves is the projection of these standing waveforms onto the bottom of the moving bed.
The results of the above infer the existence of some correlation between the geometry of the water surface and the bottom. Inspired by this, we first calculated the correlation coefficients between the water surface and water depth, and between the bottom surface and water depth, during the occurrence and development of sandbars. As a result, it was found that the water surface has a high correlation with water depth during the bar initiation stage, while the bottom has a high correlation with water depth during the bar development stage. Next, focusing on the depth-wavelength ratio obtained from the dispersion relation in the small-amplitude wave theory, we examined the depth-wavelength ratio of riverbed waves at three scales (bar, dune, and ripple) and considered the correlation between the water surface and bottom surface in these riverbed waves. As a result, it was found that the depth-wavelength ratios of rivers and channels where sandbars are formed in model experiments and in actual rivers are small, and that the assumption of shallow-water flow with a nearly uniform velocity distribution from the water surface to the bottom is valid for these flows. On the other hand, the depth-wavelength ratio of channels where dunes and ripples are formed is large, and the flow there has not a uniform distribution in the vertical direction and is not a flow that can be assumed to be shallow-water flow. In other words, it was found for the first time that in a river dominated by sandbars, the shape of sandbars on the riverbed influences the water surface, and the flow there is almost uniform in the vertical direction and can be assumed to be a shallow-water flow.
Numerical simulation has been used as the research method in recent studies of properties of sandbars. In the numerical simulation of model for flow and sediment transport, small disturbances are assumed to be a factor in the occurrence of sandbars, and small disturbances are given as initial conditions, as in stability analysis. Therefore, it is difficult to consider factors other than small disturbances in the occurrence of sandbars. In response to this problem, the authors developed a measurement system called Stream Tomography (ST), which can measure the geometrical shape of the water surface and bottom in a model experiment of the process of sandbar initiation and development with high spatial resolution and high temporal frequency, enabling the acquisition of observation big data equivalent or superior to numerical analysis. We first conducted a model experiment in which a sandbar was spontaneously occurred in a straight channel by initially spreading fine sand on a flat surface, measured the geometry of the water surface and bottom surface during the process of sandbar generation and development using ST, and analyzed the wave numbers in the flow direction and transverse direction for the water surface and bottom surface geometry. The results showed that the bottom surface remained flat immediately after the start of the model experiment, whereas the water surface had a wave train with a wave number of about 20. Next, we measured planar distribution of flow velocity by PTV using the aforementioned straight channel as a fixed bed. The results showed that the planar distribution of the velocity was equivalent to that of the water surface at the initial time of the model experiment with the moving bed, and that standing waves existed. Generally, it is said that in a straight and flat channel, the flow velocity and flow depth are spatially uniform and isentropic, but the detailed measurements revealed that standing waves exist even in a flat bed. These results suggest that one of the factors occurring sand waves is the projection of these standing waveforms onto the bottom of the moving bed.
The results of the above infer the existence of some correlation between the geometry of the water surface and the bottom. Inspired by this, we first calculated the correlation coefficients between the water surface and water depth, and between the bottom surface and water depth, during the occurrence and development of sandbars. As a result, it was found that the water surface has a high correlation with water depth during the bar initiation stage, while the bottom has a high correlation with water depth during the bar development stage. Next, focusing on the depth-wavelength ratio obtained from the dispersion relation in the small-amplitude wave theory, we examined the depth-wavelength ratio of riverbed waves at three scales (bar, dune, and ripple) and considered the correlation between the water surface and bottom surface in these riverbed waves. As a result, it was found that the depth-wavelength ratios of rivers and channels where sandbars are formed in model experiments and in actual rivers are small, and that the assumption of shallow-water flow with a nearly uniform velocity distribution from the water surface to the bottom is valid for these flows. On the other hand, the depth-wavelength ratio of channels where dunes and ripples are formed is large, and the flow there has not a uniform distribution in the vertical direction and is not a flow that can be assumed to be shallow-water flow. In other words, it was found for the first time that in a river dominated by sandbars, the shape of sandbars on the riverbed influences the water surface, and the flow there is almost uniform in the vertical direction and can be assumed to be a shallow-water flow.