*Tatsuki Kaneko1, Fumitoshi Imaizumi2, Tomoya Osada2, Raffaele Spielmann3, Shoki Takayama4
(1.The United Graduate school of Agricultural Science, Gifu University, 2.Faculty of Agriculture, Shizuoka University , 3.Chair for Engineering Geology, Geological Institute, Department of Earth Sciences, ETH Zürich, 4.Disaster Prevention Research Institute, Kyoto University)

Keywords:LiDAR, Debris flow monitoring, PIV , velocity
A debris flow is mixture of water and particles driven down a slope by gravity. To understand the debris flow dynamics, field observations of debris flows have been conducted to date. These observations have revealed that debris flows exhibit rapid changes in flow velocity and water level within one second. However, due to technical limitations, it has not been possible to simultaneously observe flow velocities and water levels with intervals of shorter than one second, so the information required to understand debris flow dynamics has been lacking. In response to this problem, advances have been made in LiDAR (Light Detection And Ranging) technology, which can measure three-dimensional shapes. One type of LiDAR, 4D (3D+time) LiDAR, can observe debris flows with the addition of sub-second time intervals, which is expected to provide new debris flow data. In this study, we aimed to understand the dynamics of debris flows by monitoring changes in water level and flow velocity at intervals of sub-second using a 4D-LiDAR automatic observation system. The study site is the Ichinosawa catchment in the Ohya landslide, central Japan. In the Ichinosawa catchment, debris flows occur several times a year by erosion of unstable sediment accumulated in the streambed during heavy rainfall. Observations were made using a 4D-LiDAR automatic observation system, which combines 4D-LiDAR, a board computer, and a program relay to automatically start observations when debris flows occur. To calculate the flow velocity from the observed 4D-LiDAR data, the surface flow velocity of debris flows was calculated by PIV (Particle Image Velocimetry). The noise included in the calculated surface flow velocities was statistically removed in the analysis. In addition, manual flow velocities were calculated by visually tracking particles and dividing the travel distance by the travel time to validate the PIV surface velocity. Furthermore, the water level of debris flows was calculated at 0.1-second intervals based on the observed 4D-LiDAR data. As a result of the observations, five debris flow events were successfully recorded using 4D-LiDAR between August 2023 and November 2024. The surface flow velocities calculated by PIV were generally consistent with the manually calculated flow velocities. In addition, the peak of the surface velocity and the peak of the flow level were shown at the same time. The surface velocity increased and decreased on simultaneously with the flow level. However, in the case of erodible bed, the surface velocity was continuously high even after the peak of flow level. Furthermore, relationship between the surface velocity and the flow level were different among debris flow events.