It is well-known that meander wavelength is empirically related to the square root of discharge. However, relations between wavelength and stream discharge are more complex than a simple two variable relationship would suggest. So far, meander wavelength has been known to be controlled by discharge, size of sediment, and slope. There has been few studies of free meanders related to the discharge in Japan, and, studies which stressed the factors controlling meander geometry are few and far between in Japan. The purpose of this paper is to analyse the wavelength of free meanders in Japan in terms of multiple regression analysis, of which independent variables are the discharge, the size of bed materials, and the valley slope. The author obtained the data extensively from the major rivers in Japan, for which discharge records were available for more than 10 vears. Since natural bankfull discharge is difficult to measure in Japan, the discharge of 1.58-year flood, which is the most probable annual flood, and also the channel-forming discharge, was used as the discharge variable. The bankfull discharge, the median diameter of bed materials, and the valley slope were selected as independent variables. On the basis of these data, a multiple regression analysis of meander wavelength was conducted stepwise. At the first step, the bankfull discharge Ql'58 Was the first independent variable being in relation to the wavelength Lm as follows; Lm = 1480 + 0.597 Ql'58 """' """"(1) Testing the regression equation by the value of F, it was found that the equation was significant at 1% Ievel of significance. On the other hand, the values of F of the variables other than the discharge, revealed that they were not significant at 1% Ievel of siguificance. After all, the wavelength of free meanders in Japan turned out to be dependent strongly on the discharge. Other factors do not contribute to the wavelength significantly. The equation (1) shows that the meander wavelength can be expressed as a linear function of the bankfull discharge, and not as a power function. The previous result obtained by the present author for 36 points in Japan, is expressed as follows; Lm = 43.9 Qmax0.53 ...... ........(2) where Lm is wavelength, and Qmax is decennary arithmetic mean of maximum discharge in a year. The equation (2) closely resembles that obtained by Dury for 133 paired observations. The reason why the equation (1)obtained from present study does not show power functional relation, may be mainly due to the fact that the data used are limited. The value of squared correlation coefficient of the equation (1) is as low as 0.596. Free meanders in Japan may have been influenced by factors other than those adopted here. We can not overlook the human interference in rivers of Japan.
抄録全体を表示