5:15 PM - 6:45 PM
[HGM03-P05] Study for the method to distinguish between mainstream and tributary gravels of fluvial terraces based on roundness using image analysis
Keywords:Doushi River, Terrace gravel, Roundness, Fluvial terrace, Image analysis
Because rivers frequently join in upper, it is reported that terrace deposits of mainstream and tributary filled upper river valleys in the Last Glacial period (e.g. Hirakawa and Ono, 1974; Koiwa, 2005; Takahashi and Sugai, 2018). As a result, mainstream and tributary terraces often coexist in an upper river (Larson et al., 2015). It is important to clarify the provenance (mainstream or tributary) of terrace deposits based on sedimentological characteristics for reconstruct the sediment budget, fluvial deposition and erosion, and fluvial terrace development in river basins.
Shape analysis of clasts is one of the useful methods to investigate the source of clast such as terrace deposits. Most of the previous studies on roundness used qualitative and semi-quantitative methods based on visual judgment and measured limited number and size of clasts. In addition, the finer clasts of fluvial sediments tend to lower roundness even in braided rivers of mainstream due to fracturing (e.g. Utsugawa and Shirai, 2019). It is necessary to clarify the most suitable grain-size to distinguish between mainstream and tributary deposits and to do sampling and analysis efficiently. Recently, a particle shape analysis method using computer image analysis has been established (Zheng and Hryciw, 2015), and has made it possible to obtain quantitative particle shape measurement data with unprecedented accuracy and efficiency, even for finer clastic material of terrace deposits. In this study, the image shape analysis of roundness was applied to the clasts of the present river and terrace gravel in the Doushi River (mainstream) and its tributary, Kajino Stream in Sagami River basin, central Japan. We compared the roundness distribution of gravels in the present riverbed and terrace gravels in the mainstream and tributary and examined the most useful grain sizes to distinguish mainstream and tributary deposits based on roundness.
There was no significant difference in roundness between the present mainstream riverbed gravels and the mainstream terrace gravels for all grain sizes in 2–32 mm. However, comparing the present riverbed gravels from tributary and sub-tributary, the sub-tributary gravel has lower roundness than tributary gravel. Also, all samples of tributary terrace gravels sampled from several horizons in this study have similar roundness distribution to that of the sub-tributary. This reflects the difference in river slopes between the tributary and sub-tributary, tributary with mean slopes below 8%, which is the debris-flow stopping gradient, are traction tributary, while sub-traction with mean slopes above 8% are considered to be debris-flow tributary (Shimazu, 1991).
These suggest that the sampled tributary terrace gravels were supplied by a debris-flow from the sub-tributary. Comparing the roundness distribution of mainstream and tributary gravels in the present rivers, the mean roundness of the gravels was clearly different only in 16–32 mm grain size. On the other hand, the mean roundness was distinguishable between the mainstream and the sub-tributary by larger than 4–8 mm. These differences may be caused by the difference in transportation processes between the tributary and the sub-tributary. The difference between samples with different transportation processes (debris-flow or traction) and lithology was significant even for fine grains, while the difference between samples with the same transportation processes but different lithology were significant only for the coarse grains. The results imply that the roundness of gravel clasts in rivers varies strongly depending on the transportation processes of that place in the river, and that differences in the roundness distribution brought about by differences in lithology and basin area are limited.
Shape analysis of clasts is one of the useful methods to investigate the source of clast such as terrace deposits. Most of the previous studies on roundness used qualitative and semi-quantitative methods based on visual judgment and measured limited number and size of clasts. In addition, the finer clasts of fluvial sediments tend to lower roundness even in braided rivers of mainstream due to fracturing (e.g. Utsugawa and Shirai, 2019). It is necessary to clarify the most suitable grain-size to distinguish between mainstream and tributary deposits and to do sampling and analysis efficiently. Recently, a particle shape analysis method using computer image analysis has been established (Zheng and Hryciw, 2015), and has made it possible to obtain quantitative particle shape measurement data with unprecedented accuracy and efficiency, even for finer clastic material of terrace deposits. In this study, the image shape analysis of roundness was applied to the clasts of the present river and terrace gravel in the Doushi River (mainstream) and its tributary, Kajino Stream in Sagami River basin, central Japan. We compared the roundness distribution of gravels in the present riverbed and terrace gravels in the mainstream and tributary and examined the most useful grain sizes to distinguish mainstream and tributary deposits based on roundness.
There was no significant difference in roundness between the present mainstream riverbed gravels and the mainstream terrace gravels for all grain sizes in 2–32 mm. However, comparing the present riverbed gravels from tributary and sub-tributary, the sub-tributary gravel has lower roundness than tributary gravel. Also, all samples of tributary terrace gravels sampled from several horizons in this study have similar roundness distribution to that of the sub-tributary. This reflects the difference in river slopes between the tributary and sub-tributary, tributary with mean slopes below 8%, which is the debris-flow stopping gradient, are traction tributary, while sub-traction with mean slopes above 8% are considered to be debris-flow tributary (Shimazu, 1991).
These suggest that the sampled tributary terrace gravels were supplied by a debris-flow from the sub-tributary. Comparing the roundness distribution of mainstream and tributary gravels in the present rivers, the mean roundness of the gravels was clearly different only in 16–32 mm grain size. On the other hand, the mean roundness was distinguishable between the mainstream and the sub-tributary by larger than 4–8 mm. These differences may be caused by the difference in transportation processes between the tributary and the sub-tributary. The difference between samples with different transportation processes (debris-flow or traction) and lithology was significant even for fine grains, while the difference between samples with the same transportation processes but different lithology were significant only for the coarse grains. The results imply that the roundness of gravel clasts in rivers varies strongly depending on the transportation processes of that place in the river, and that differences in the roundness distribution brought about by differences in lithology and basin area are limited.