11:00 AM - 1:00 PM
[HGM03-P05] Roundness change of gravel particle along the rivers based on image analysis: a case study of Joganji, Sagami and Shimanto rivers
Keywords:Roundness, Shimanto River, Joganji River, Sagami River
Many kinds of particle shape parameters have been developed (Blott and Pye, 2008). Among them, the degree of roundness is often used to describe shapes such as gravel. In this study, we use Wadell’s roundness (hereinafter R) defined by Wadell (1932), and R is a value from 0 to 1. In Japan, R has been measured for gravel on rivers and beaches by Nakayama (1954, 1965) and Nakayama and Miura (1964). However, in these studies, only the mean value is used, and there is no discussion using the distribution of the R histogram. Recently, changes in particle shape and mass during the particle transport process were discussed in more detail using experiments and numerical modeling (e.g., Domokos et al., 2014). However, sufficient datasets in the field are not obtained, and comparison with these models and experimental results is needed in the future. In this study, we aim to obtain R in large quantities and with high precision based on image analysis and show the change in R from upstream to downstream.
We conducted analyses of gravel particles along the Joganji, Sagami, and the Shimanto rivers which have different gradients. We set the interval of sampling sites to 5–30 km and chose sampling points where artificial effects were less. The analyzed particle size was 2–64 mm, and samples were taken every 1 phi. We calculated R using the code by Zheng and Hryciw (2015) and followed Ishimura and Yamada (2019) for other settings.
In the Shimanto River, the lowest and largest mean values were obtained at the point about 4 km from the upstream end and on the coast, respectively. On the other hand, almost the same mean values were shown in the section between them, showing a tendency to gradually increase toward the downstream. Although the skewness varies a little, the larger the particle size, the more upstream side the skewness changes from a positive value to a negative one.
In the Sagami River, gravel could not be collected in a section in the mountainous areas due to the dam lake, however, we could collect the gravels in the most upstream and downstream areas. The tendency obtained in the river is similar to the Shimanto River; a rapid increase in the mean value was observed between the point 4 km from the upstream end and the next 11 km point, and a gradual increase in the mean value was observed on the downstream side. Although the skewness varies a little, it shows a decreasing tendency toward the downstream as a whole.
In the Joganji River, the sample of the most upstream part could not be collected, so the change from the mountainous area to the river mouth is described. No abrupt change in the mean value was observed, and the mean value tended to increase gradually toward the downstream. The skewness is not as clear as that of the Shimanto River, but the larger the particle size, the smaller the value.
From the above results from three rivers, the change in R toward the downstream shows a tendency to increase rapidly at the most upstream part and then gradually increase. This tendency reflects the rapid increase in roundness/circularity at the beginning of the attrition process of particles, as shown by Krumbein (1941) and Novák-Szabó et al. (2018). Although the mean value of R does not change significantly, the skewness changes from positive to negative toward the downstream. This indicates that the skewness reflects the degree of saturation of R.
The mean value of R saturated at a certain value in the Shimanto River, but the mean value of R increased further on the beach adjacent to the river mouth, which may indicate the difference in the attrition process between the river and the coast. The mean value of saturated R in a river differs for each river and each particle size and shows a value of 0.4–0.6. In all three rivers, the larger the particle size, the larger the value of saturated R.
We conducted analyses of gravel particles along the Joganji, Sagami, and the Shimanto rivers which have different gradients. We set the interval of sampling sites to 5–30 km and chose sampling points where artificial effects were less. The analyzed particle size was 2–64 mm, and samples were taken every 1 phi. We calculated R using the code by Zheng and Hryciw (2015) and followed Ishimura and Yamada (2019) for other settings.
In the Shimanto River, the lowest and largest mean values were obtained at the point about 4 km from the upstream end and on the coast, respectively. On the other hand, almost the same mean values were shown in the section between them, showing a tendency to gradually increase toward the downstream. Although the skewness varies a little, the larger the particle size, the more upstream side the skewness changes from a positive value to a negative one.
In the Sagami River, gravel could not be collected in a section in the mountainous areas due to the dam lake, however, we could collect the gravels in the most upstream and downstream areas. The tendency obtained in the river is similar to the Shimanto River; a rapid increase in the mean value was observed between the point 4 km from the upstream end and the next 11 km point, and a gradual increase in the mean value was observed on the downstream side. Although the skewness varies a little, it shows a decreasing tendency toward the downstream as a whole.
In the Joganji River, the sample of the most upstream part could not be collected, so the change from the mountainous area to the river mouth is described. No abrupt change in the mean value was observed, and the mean value tended to increase gradually toward the downstream. The skewness is not as clear as that of the Shimanto River, but the larger the particle size, the smaller the value.
From the above results from three rivers, the change in R toward the downstream shows a tendency to increase rapidly at the most upstream part and then gradually increase. This tendency reflects the rapid increase in roundness/circularity at the beginning of the attrition process of particles, as shown by Krumbein (1941) and Novák-Szabó et al. (2018). Although the mean value of R does not change significantly, the skewness changes from positive to negative toward the downstream. This indicates that the skewness reflects the degree of saturation of R.
The mean value of R saturated at a certain value in the Shimanto River, but the mean value of R increased further on the beach adjacent to the river mouth, which may indicate the difference in the attrition process between the river and the coast. The mean value of saturated R in a river differs for each river and each particle size and shows a value of 0.4–0.6. In all three rivers, the larger the particle size, the larger the value of saturated R.