Flood discharge observations in Japan are shifting from the conventional float-based methods to unmanned techniques such as radio-wave current meters. These approaches differ fundamentally in their measurement principles: the former is based on a Lagrangian framework, whereas the latter relies on a Eulerian framework. In this study, surface velocity fields obtained using particle image velocimetry (PIV) were used to track virtual tracers and derive Lagrangian surface velocities, providing a basis for examining the characteristics of Lagrangian and Eulerian measurements in an actual river under flood conditions. The uncertainties associated with the two frameworks were quantitatively compared, and the principal sources of uncertainty in Lagrangian measurements were identified. To achieve accurate discharge observation based on Eulerian measurements, the influences of measurement duration, subsection configuration, and vertical velocity distribution were investigated. The results suggest that measuring many points over a short duration is more effective than measuring a few points over a long duration. In a fixed-point measurement of subsurface velocity, a velocity dip was observed. Furthermore, the results quantitatively demonstrate the effects of bridge-pier wakes on the required averaging time and subsection configuration, highlighting the practical advantage of conducting observations on the upstream side of bridges.
Kudo et al. (Thu,) studied this question.