As global warming progresses, there is concern that heavy rainfall will become more severe, and it is hoped that meteorological control will help to reduce heavy rainfall. In this study, Large-Eddy-Simulation (LES) was used to reproduce the 2008 Guerrilla heavy rainfall at Toga River, Kobe, Japan, and the effect of controlling the near-surface wind field to weaken vortex tubes, which are regarded as an important factor in the development of the Guerrilla heavy rainfall, was investigated. The results showed that the LES can represent complex cases of guerrilla heavy rainfall such as the case of Toga River, and that the wind field manipulation can weaken the maximum rainfall intensity by about 27%. The suppression mechanism is suggested to be the suppression of updrafts due to the weakening of vortex tubes, misalignment between the location of the temperature anomaly that causes updraft and the location of the airflow convergence, and the reduction in the amount of water vapor uptake. In conclusion, a meteorological control method to suppress heavy rainfall by manipulating the wind velocity field from meteorological perspective.

 From the aspect of hydrodynamic, when the atmosphere is unstable, the flow tends to overcome the topography vertically, while when the atmosphere is stable, it tends to avoid the topography and go around it horizontally. In this study, we hypothesized that the more stable the atmosphere is, the more water vapor will be transported to inland, and conducted the sensitivity experiments on atmospheric stability in order to verify the hypothesis. Sensitivity experiments were conducted for the 2012 Kameoka heavy rainfall event, which was caused by the abundant water vapor supplied through the Osaka Bay, then we revealed that the amount of water vapor supplied from the south to inland was greater when the atmosphere was more stable. Furthermore, a cluster analysis of the past atmospheric field was conducted to verify the result of sensitivity experiment results, and it was shown that more water vapor tended to flow inland when the past actual atmosphere was stable. This study provides very important fundamental knowledge for considering the relationship between heavy rainfall and global warming, which has the effect of stabilizing the atmosphere.

 Kyoto is one of the most flood-prone cities in the Kansai region. It is essential to understand the precipitation system that causes heavy rainfall for disaster prevention and mitigation. In this study, we analyzed the spatiotemporal structure and environmental field of the precipitation system inducing the heavy rainfall event in Kyoto city by using radar AMeDAS rainfall data and GPV-MSM reanalysis data. We defined the two thresholds for extracting the heavy rainfall event in different temporal scales; (1) daily maximum 1-hour precipitation exceeds 20 mm and (2) daily maximum 10-minute precipitation exceeds 10 mm. The events meeting only (1) threshold are characterized by a line-shaped precipitation system originating from Mt. Rokko. These are associated with a developed low-pressure system with a trough at 850hPa. In contrast, the events meeting only (2) threshold are localized precipitation systems over Kyoto city, suggesting convective precipitation with warmer surface temperature in the Kansai area due to the expansion of the Pacific high to the west.

 In this study, we utilized a self-organizing map to analyze the relationship between the meteorological fields across the entire Kyushu region, where is increasing risk of heavy rainfall in the future, and the rainfall patterns in the first-class river basins. River basins selected in this study include the gala, Rokkaku-kawa, Ono-gawa, Shira-kawa, Kuma-gawk, Oyodo-gawa, and Sendai-gawa. As a result, it was found that the basins in the western part of northern Kyushu (Chikugo-gawa, Rokkaku-gawa, and Shira-kawa) were predominantly affected by heavy rainfall from frontal systems. Heavy rainfall events that occurred in the basins of the western part of southern Kyushu (Kuma-gawa and Sendai-gawa) were characterized by typhoons as well as frontal systems. On the other hand, in the eastern basins (Ono-gawa and Oyodo-gawa), typhoon-induced heavy rainfall was highly prevalent. Moreover, the meteorological pattern on 14 August 2021, which induced heavy rainfall in Saga Prefecture, was diagnosed by using the self-organizing map constructed in this study. As a result, it was found that this meteorological pattern indicates a high risk of heavy rainfall in Chikugo-gawa and Rokkaku-gawa river basins.

 The purpose of this research is to analyze environmental conditions of stagnant frontal line-shaped convective systems, which trigger heavy rainfall disasters in Baiu season in Japan, using past events and considering spatial and temporal characteristics. As a result, type-A, which is large scale and occurs near convergence of Baiu front, occurs by large convergence of Baiu front, so the environmental condition of type-A is not necessarily idealized. On the other hand, type-B, which is small scale and occurs more than 100km away from Baiu front, occurs at more limited and idealized environment than type-A. We consider that strictly environmental conditions contribute to self-organized development of the first convective cloud. We also showed that requirements for occurrence of the first cloud are convective instable, large CAPE (Convective Available Potential Energy), and big change of direction of wind. Moreover, our results bring that the constant ratio of buoyancy to turbulence is important for occurrence and development of line-shaped convective systems.

 Convective systems that form cloud streets during the winter season when the monsoon from the northwest prevails over Japan. They consist of cloud regions where water vapor condenses and clouds are formed in the updraft region of convection driven by the temperature difference between the warm ocean and the cold atmosphere, and non-cloud regions where no clouds exist in the downdraft region of convection. In this study, we extracted cloud streets that appeared near Ishikari Bay from November to March in 2015-2020, and showed that the difference in water vapor between cloud and non-cloud areas was on average 1.1 times greater in cloud areas with a standard deviation of 17%. Furthermore, we explained that the difference of both in water vapor is caused not only by the water vapor budget but also by changes in water vapor capacity due to thermodynamic effects, and proposed a method for estimating the contribution of water vapor condensation inside clouds and the effect of the uptake of surrounding air.

 In this study, model tests were conducted focusing on the amount of water leakage in the levee in relation to the external water level, and a quantitative understanding of the risk of piping failure was attempted based on the relationship between the amount of water leakage and the degree of piping propagation after correcting for differences in soil and hydraulic conditions. The results showed that a limit range exists for the cumulative amount of leakage from the time of leakage to piping failure. From there, a new concept of piping failure curve (fatigue curve of levee strength versus time of action of the external water level) was developed. Future field flow measurements are desired for application to actual levees.

 To prevent and reduce house damage due to flooding, it is important to understand the quantitative evaluation for house damage under a wide variety of hydraulic conditions. This study aims to clarify the relationship house damage and hydraulic properties under the temporal variation of the breach width. The target was the flooding in the Kinu River with the levee breach at 21 kp, in which the width varied temporally. For this, we conducted a coupled numerical analysis for river flow and flooding flow. The results indicated that the flow velocity and inundation depth around the point of levee breach varied appreciably with the breach widths. The velocity and inundation depth at most of collapsed houses were larger than 2 m/s and 2 m, respectively, which are a threshold for house collapse. On the other hand, there were completely collapsed houses with a small flood depth below the collapse limit, suggesting the importance of scouring the ground around houses.

 Huge floods due to heavy rain have occurred in the world and caused severe house damages. It is therefore necessary to evaluate the relation between flood flow and house damage. The purpose of this paper is to evaluate the relation between flood flow and house damage in the Chikuma River flood caused by the Typhoon Hagibis in 2019. For this, a combined analysis of river flow and inundation flow was conducted. The calculated results showed that the flood flow from the breach point was separated into three directions, and the flow direction of high fluid force was comparable to the band profile of the house loss. However, the fluid force and velocity on the houses away from the breach point was small, suggesting further improvement of the numerical accuracy using fine computational grid.

 This study constructs a model that combines a distributed rainfall runoff flood inundation analysis and slope stability analysis for accurately analyzing river floods and landslide caused by heavy rainfall. The past rainfall data of the weather radar MP-PAWE installed at Saitama University was used for validating the model. The 10m mesh distributed rainfall-runoff flood inundation analysis and separating the floodwater by rivers or rain source, enabled a more detailed analysis of inundation patterns in tributary rivers. Further improvement in accuracy can be expected by modeling drainage channels, sewers, reservoirs, etc. The actual slope failure locations coincided with the locations where the factor of safety was reduced in the slope stability analysis. In order to accurately predict the collapse time, it is necessary to consider the vertical infiltration process of rainwater and to devise a method for setting parameters.

 Floods in urban areas cause inundation damage. In addition, drifting objects such as vehicles can cause collision damage to buildings and increase the inundation area by damming up the flowing water.

 In this study, a numerical model was developed to predict influences of drifting vehicles on distribution of inundated areas. Computed results obtained under the straight channel condition show that the critical condition of the movement of stationary object can be calculated and that the presence of drifting objects in flowing water causes increases of water depth at the upstream section. It was also shown that the presence of drifting objects in flooded area changes the extent and depth of the inundated area by obstructing the spreading of the inundation flow.

 In recent years, typhoons and fronts have caused frequent large-scale torrential rains, and localized torrential rains can cause damage before warnings are issued. In this study, the number of vehicles calculated from cross-sectional traffic volume was incorporated into S-uiPS, a sophisticated flooding prediction method, for the Shibaura Sewage Treatment District in the 23 wards of Tokyo, and flooding prediction calculations were conducted for an assumed maximum-scale rainfall event reflecting traffic congestion. As a result, it was found that the inundation depth increases on arterial roads and surrounding roads. In addition, a route search system that can avoid flooding from populated areas to hospitals in the same area was studied. By linking the system with inundation analysis, routes that are passable even during heavy rainfall were obtained, and shorter routes were obtained when the time increments were reduced.

 After the disappearance of sediment, bedrock exposure and erosion progressed, and knickpoint was formed at Toyohira River Oiranbuchi. In this study, we attempted to elucidate the erosion process of bedrock rivers with change in river width and slope using a numerical model that considers bedrock erosion. Numerical experiments were performed by changing the river width and discharge. As a result, it was suggested that the change in the river width around the knickpoint increased the amount of sediment transport and it made the amount of bedrock erosion, which promoted the retreat of the knickpoint. In addition, the dry-wet weathering flattened the upstream bedrock and retreated the entire steep slope, suggesting that the effect of the change in river width on the knickpoint retreat was maintained.

 The formation of sand waves is an important issue in river engineering, since they affect the flow and sediment transport, and their resistance to the flow changes during the formation and transition processes of dunes. In this study, a three-dimensional flow model was combined with a sediment transport model that takes into account the effect of pressure gradients near a riverbed and applied to the formation process of dunes, and the validity of the model was verified. The numerical model reasonably reproduces the formation process in which the relatively regular waves generate in the early stage change to the dunes, and the upstream dunes catch up with and merge with the downstream dunes to form large-scale dunes. In addition, the flow characteristics before and after the merge of dunes are examined from the numerical results.

 Predicting changes in bar wave height is important from the perspective of channel management. In this study, we perform theoretical and numerical analyses to discuss the bar growth rate relative to the flow discharge and current bar wave height. In addition, we propose a simple model to predict wave height development for a channel with flood plain limited by levees. We develop the model based on the numerical results, referencing the bar growth rate equation used in the theoretical analysis. Then, we input a hydro with flood control by the dam and a hydro without flood control into the model. As a result, the former case showed a continuous increase in wave height, whereas the latter case showed a decay of wave height at near the peak flow discharge, so that the final wave height did not increase significantly. This suggests that dam flood control may have a significant effect on the increase in bar wave height.

 Many studies have been conducted on spontaneously formed sandbars in rivers where the bed material consists of sand and gravel, but the physical mechanism of their formation has not yet been clarified. In this study, we measured the bottom and water surface of a single row of sandbars at the time of their formation using ST, calculated the flow velocity by plane two-dimensional analysis, and calculated Pe, which is the ratio of advection to diffusion. We found that there is a close relationship between Pe and the geometry of the bottom surface, and that the histogram of the x-component of Pe is bimodal. In particular, the distribution of the histogram changed at a constant value in the early stage of sandbar formation, regardless of the conditions. These results suggest that the large Pe during the formation process of sandbars is due to the fact that the advection of sandbars outweighs the diffusion of sandbars, and that the wave height of riverbed waves develops due to this equilibrium relationship.

 Fluvial bar movements are an important subject for sediment hydraulics and river managements. In this study, fluvial bar movements at a river confluence of Tone River and Karasu River, which caused the burring of water level gauge of Yatta-jima WLGS in 2011 due to the elongation of leftbank bar at the confluence, were investigated by using long-term analyses of aerophotos and hydrological data, and by conducting movable bed experiments. In conclusions, the elongation of leftbank bar started almost from about 2001. It was suggested that this elongation was caused by the change of river discharge ratio due to precipitation changes at upstream river basins. On the basis of this suggestion, movable bed experiments and visualization by using fluorescent sands by using a simple open-channel flume of Y-shaped confluence were conducted by varying the river discharge ratio. From experimental results, it was revealed that floods due to the Tone impact, which indicated river discharges of Tone River were larger than those of Karasu River, caused the expand and elongation of leftbank bar at river confluence and those due to the Karasu impact after the Tone impact flood caused the straight-shaped elongation of leftbank bar, which was almost the same shape with field bar as shown in final aerophoto in 2011.

 The flow cell type ultrasonic attenuation spectrometer was the first suspended sediment measuring system to be installed in a river, in 2017. Although wash loads had been measured in previous experiments, there were problems with the sampling method and the flow cell structure for the measurement of sediment containing particles larger than fine-grained and medium-grained sand. In this study we increased the sample volume, improved the particle size distribution determination method, and measured suspended sediment during snowmelt runoff using an ultrasonic attenuation spectrometer with an improved flow cell. We confirmed that this spectrometer could measure the concentration and the particle size distribution of suspended sediment with diameters of 2 mm or less and wash load. Using this data, it was possible to estimate total load by particle size class.

 Treetop detection was conducted in a dense bamboo forest using Local Maximum Filtering (LMF) with digital surface models (DSMs), which were generated using airborne laser scanning data and UAV images processed with Structure-from-Motion/Multi-View-Stereo (UAV-SfM). Subsequently, LMF was applied to both DSMs and the optimal Window Size (WS) parameter for LMF was examined. The results showed that the optimal WS was smaller for the UAV-derived DSM (DSM-U) than for the ALS-derived DSM (DSM-A). This difference was attributed to smoother elevation roughness in the DSM-U than in the DSM-A. These results suggest that the types of DSMs generated by ALS or UAV-SfM data affect the application results of treetop detection. Therefore, it is important that the WS should be set considering the characteristics of the DSMs.

 Space-Time Image Velocimetry (STIV) is a flow visualization method specified to measure the surface velocity distribution of rivers by tracing the surface waves that emerge on the river surface due to the turbulence of the water flow. Surface tracers are distributed for measuring low water flow in which the surface waves are indistinct. Surface tracers gather to the converging area of surface velocity, resulting in inhomogeneous tracer density, which causes problems in statistical analysis. The temperature difference between the river bed, the water flow and the atmosphere and the turbulence cause slight variations in the surface water temperature in the river flow. The turbulence characteristics were analyzed utilizing image acquisition by a mid-wavelength infrared camera and the unique feature of STIV, and then obtained flow characteristics were compared with the results obtained with PIV. The results suggest that the proposed method can more accurately acquire turbulent statistics.

 In the planning and design of rivers, it is necessary to know the flow rate during floods and the depth of the foundation depth at which the structure can maintain stability during floods. For this purpose, it is essential to understand the riverbed fluctuations during floods, but methods for quantifying the phenomena in flowing water during floods have not yet been established. Surface velocity is one of the hydraulic quantities that can be obtained even in turbid water during floods. Theoretically, the Mass-Consistent model can be applied to calculate the flow depth from the known surface velocity, and the river bed can be determined through the flow depth. In this study, we developed a method for estimating two-dimensional stream depth based on a continuity equation using a planar velocity distribution as an input value. Numerical experiments have shown that the method can be estimated the flow depth distribution in a flow simulating sandbars, and that the estimation accuracy improves when the weight coefficient is used as a spatial variable. The noise in the input values is tolerated to the same degree as the amplitude of the flow depth.

 In this study, we investigated the uses of multispectral cameras and satellite data with the aim of developing river turbidity measurement methods for various scales. As a result, it was found that the turbidity estimation using the same turbidity estimation equations is possible because the turbidity reflectances obtained from the multispectral cameras and satellite data are almost the same. Furthermore, the near-infrared wavelengths showed the possibility of constructing a turbidity estimation formula applicable to actual rivers from experimental data at near-infrared wavelengths. On the other hand, it was found that the field observation data is necessary because the red wavelength is affected by the water depth.

 This study aims to obtain river discharge with spatial distribution of velocity by video image analysis. Target event is the experiment at the Chiyoda Experimental Flume in June 2021. First, Particle Image Velocimetry with geometric transformation by information of angle of camera and distance from camera to water surface was conducted. The observed velocity by PIV was compared with the velocity by a radio wave current meter, which showed a reasonable accuracy. Secondly, discharge calculated with the velocity distribution by PIV, observed water level and estimated river bed elevation by hydraulic resistance theory showed a good agreement with the actual discharge in the experimental flume.

 In order to discuss the hydraulic properties in the actual river, the authors have developed an observation method. The main target of the flow is the large size flood, which recently take place. Using the video camera mounted on UAV, PIV for velocity and SfM for water surface are implemented in order to obtain the total head. The flow generated at Chiyoda experimental facility is used for the measurements. From the observed results, micro-bed forms were recognized. The water depth was estimated from the hydraulic resistance. The energy loss were also estimated by using the energy slope along the streamline, which are above the micro-scale bed form, and the river protection works.

 The increase in marine debris is concerned to have an irreversible negative impact on ecosystems. Riverine litter, a major source of marine debris, is causing damage to the landscape and adverse effects on the ecosystem of rivers themselves. This study con-ducted field surveys to materials and quantity of dumped and drifted litter on riverbanks and litter by a screen in an upstream channel in the Yamato River. From the field survey in the riverbank of the Yamato River, the correlation coefficient between the weight of litter and coverage is 0.82. The litter weight per 1 % coverage ratio was estimated to be 26.8 g/m2. The weight of litter can be estimated using the coverage ratio, but the value is affected by types of litter and is more overestimated when the sampled litter contains more plastic and paper wastes. The analysis of drifting litter trapped by a silt screen deployed in the channel in the Saho river, the upstream of the Yamato River, during dredging maintenance clarified that about 43.1 % of the litter was PET bottles on weight basis, and it was estimated that about 112 kg of litter would have passed through the river during a period of 54 days. The best-by date of plastic bottles and cans and the estimated date of manufacture by subtracting the best before period shows that about 91.2% of the plastic bottles and cans have been manufactured within one year before the screening.

 The degradation degree regarding the molecular structure, surface roughness, and weight of a drinking bottle’s lib was evaluated based on the UV exposure experiment. As the lib weight decreased, the surface roughness increased. Furthermore, a significant correlation was found between the surface roughness and the ratio of the degraded lib weight to its initial weight, suggesting nano-sized plastic particles were generated from the lib surface. Therefore, a generalized linear model was constructed with UV exposure as the explanatory variable and the rate of mass loss as the response variable. Using the GLM, the annual generation of microplastics from the libs of drinking bottles littered in the river basin was estimated at 22.6 kg. This is little contribution (less than 0.01%), compared with the annual plastic waste flux from Japan, but further research on the fragmentation of macro-plastics littered in a river basin will be needed.

 Using the deep learning model, YOLOv5, we constructed a system to continuously detect floating debris in rivers from water surface images captured by fixed-point cameras. This system gave good detection in accuracy results. We conducted observations of floating debris using multiple fixed-point cameras and the developed system in the Hirano River, an urban tidal river in Osaka, and analyzed the spatiotemporal behavior characteristics of floating debris. The behavior of floating debris was mainly controlled by tidal currents, and the accumulation was observed in the upper reaches of the river, where the flow was stagnant at high tide. The density of floating debris tended to be larger in the upstream of the river and smaller in the downstream, due to the material diffusion caused by the fluctuating component of the tidal flow was larger in the downstream than in the upstream. In the Hirano River, a large amount of floating debris was carried downstream even during sunny days, and it was considered that the amount of debris directly dumped on the water surface increases due to easy human access along the river.

 Random Forest (RF), a machine learning method, was applied to classify phytoplankton composition using a multiple wavelength fluorescence spectrometer (MFL). First, the validity of the RF application was verified using pseudo-observational data obtained from a culture strain that simulates the actual site. Next, the applicability of RF was examined using observation data obtained at a dam reservoir. Four composition groups (Brown: diatoms, dinoflagellates, chrysophytes; Green: green algae, chlorophytes; Mixed: cryptophytes; Blue: cyanobacteria) were selected based on their fluorescence characteristics. As a result, relatively good classification accuracy was obtained for Brown and Green, and a certain degree of accuracy was obtained for Blue. However, the Mixed group could not perform with reasonable accuracy. In addition, when the RF learning model obtained by the filed data in a dam reservoir was applied to spatial MFL data acquired at the same dam reservoir, characteristic distributions of Brown and Green were obtained, indicating the usefulness of MFL.

 Effects of Egeria densa on aquatic ecosystems are major concerns in Japan. It is important to efficiently identify river segments in the entire watershed where the E. densa can be overgrowth. Using environmental DNA (eDNA) analyses, we examined the distribution and abundance of E. densa and the factors that could affect the overgrowth of the target species in the Gonokawa River. In addition, the potential maps of the suitable habitats for E. densa overgrowth were created based on the results. Our results indicated that the riverbed slope and low water temperature time were significantly and negatively related to the eDNA concentrations of E. densa in the study river basin. In addition, based on the potential maps created with the statistical results, the river segments suitable for the E. densa overgrowth were identified in the study river basin.

 Water lettuce, Pistia stratiotes is widespread in tropical and sub-tropical regions, and the excessive growth of water lettuce causes ecological and socio-economic problems. The problems have also been reported in the western part of Japan, and the possibility has been pointed out that climate warming may prompt the excessive growth of water lettuce. Therefore, a model for growth process of Pistia stratiotes populations was established and validated for two observational data. Then, some efficient ways to control and reduce the invasive macrophyte were discussed.

 The Editorial Sub-committee of Journal of JSCE (Hydraulic Engineering) had published the old version of the following manuscript and therefore retracted it.
 The correct version of this paper is published in the Journal of JSCE, Vol.80, No.16, 23-16197, 2024.

 Three-dimensional wind fields that can be estimated from meteorological radar data are useful for short-time rainfall prediction such as localized torrential rainfall. In this study, horizontal and vertical wind fields were estimated by the VVP method and the Dual Doppler analysis using three-dimensional high-density observation data obtained by dualpol phased-array weather radar. The results showed that the frequency of weak winds decreases and the frequency of strong winds increases as the elevation density of the input data increases. The estimated wind speeds also showed a smaller scatter. Furthermore, the estimated vertical winds from the Dual Doppler analysis showed less variation than those estimated by the VVP method.

 In Japan, there are many steep terrains, and the mountains enhance the rainfall resulting in much of total rainfall amount. Due to the effect of orographic rainfall, there is sometimes a discrepancy between radar-observed rainfall and ground-observed rainfall. Accurate estimation of ground rainfall is very important from the viewpoint of disaster prevention. In this study, in order to improve the accuracy of the ground rainfall estimation, I have developed a method to estimate the capture rate in real time from the radar information, and the values of the capture rate obtained by the method are used for the ground rainfall estimation. As a result, the capture rate was calculated reliably. It is also shown that the method can estimate the ground rainfall with the same or better accuracy than the conventional prediction method, even when using the capture rate obtained from the real-time radar information.

 Usually, in deriving rain rates from rain gauge observations, observed rain amount is divided by the observation time duration (called simple method). In the case of tipping bucket, the simple method has poor performance for shorter time duration. The authors developed a new method called interpolation method, where the time of tipping is estimated in statistical manner, accumurated rain amount is interpolated in time and the rain rate is derived. In this study, using a disdrometer, rain gauge observation data is simulated. Then, the simple method and interpolation method are applied to the simulation data, and the accuracy is evaulated. It is shown that the interpolation method is more accurate than the simple method particularly for shorter time duration. In the interpolation method, the accuracy becomes higher with shorter rain gauge recording period. The method to derive 1-hour rain rate from hourly rain gauge observations has been improved further by modifying the method to estimate the time of tipping.

 Horizontal micro-droplet fluxes above forest canopies were assumed to predominantly contribute to the rainfall interception process. The product of wind speed and rainfall intensity was stochastically modelled as horizontal micro-droplet fluxes on the basis of turbulent diffusion theory. The horizontal micro-droplet fluxes were applied to a multiple regression model as independent variables. The rainfall interception and meteorological data observed in two forested catchments in Tokushima were applied to the multiple regresion model. The results showed that the proposed multiple regression model described the hourly processes of rainfall interception well.

 Numerical methods are highly effective for large-scale and violent fluid phenomena such as tsunamis and floods. Among them, the Lattice Boltzmann Method is an explicit computational approach with high parallelization efficiency. In recent years, the combined use of the cumulant model and Phase-Field method has enabled highly stable analysis of multiphase fluid flows. However, there have been limited applications in the field of civil engineering. Therefore, in this study, we developed a multiphase flow numerical model based on the cumulant model and Phase-Field method specifically intended to analyze intense flow fields. As a validation of the model, we analyzed droplet oscillations in zero-gravity condition and a water column collapsing under gravity. The results showed good accuracy in representing surface tension as well as surge speed. Additionally, by conducting an analysis of water column collapse with obstacles, we confirmed accurate predictions of water level changes and wave pressure under the violent flow condition.

 Because of its simplicity, the water depth-integrated equation based on a hydrostatic assumption (longwave model) is often used to simulate river flows. On the other hand, the Boussinesq equations should be used to simulate progressive waves, such as solitary waves. In other words, the accurate representation of the non-hydrostatic components is an important aspect, and previous studies have proposed depth-integrated equations based on variational principles. Therefore, in this study, we performed simulations of progressive waves and steady & rapid flows using the longwave and Wave-Current models using the Variational principle (WCV model) and investigated their fidelity and reliability. Furthermore, we also compared the results with a three-dimensional numerical model to explore the effect of vertical vortices in the cross-sectional plane. As a result, it was found that the WCV model should be used to simulate the flow in rapid flow conditions. Additionally, in flow fields where influential vortices are formed near the water surface, it was found that a three-dimensional numerical model capable of reproducing the vertical vortices should be used instead of the water depth-integrated equation with no vorticities.

 In this study, a novel leak detection method for pipeline using pressure transient data assimilation with a numerical simulation incorporating a numerical scheme corresponding to a single leak point was developed. The accuracy of estimating the leak location and leakage volume for a single leak point was verified using this method. As a result, in both model pipeline experiments and field pipeline experiments, the estimated leak location error relative to the total length of the pipe was approximately less than 2%, and the leak volume error relative to the measured value was approximately less than 30%, indicating that this leak detection method has a practically feasible level of performance. Furthermore, the limitation of leak detection in terms of leak scale was determined to be approximately 1.70×10^-5 and 2.73×10^-3 times the pipe cross-sectional area in the model and field pipeline experiments, respectively, which corresponds to be significantly smaller than those reported in previous research, and at a similar level to them.

 For classical hydraulic jumps with an undeveloped inflow condition below a sluice gate, the magnitude and the distribution of the air concentration have been systematically investigated for a wide range of Reynolds numbers under a given inflow Froude number. The effects of Reynolds number on the air concentration distribution in the jumps are shown. The high-speed video camera images show that the amount of air bubbles into the jumps decreases as the value of Reynolds number decreases if the air-concentration ratios are affected by the Reynolds number.

 It has been pointed out that overtopping is the main cause of the recent years high number of levee failures. The Ministry of Land, Infrastructure, Transport and Tourism (MLIT) has launched the Technical Study Group on Strengthening River Levees with the aim of developing " Persistent River Levees " against water overtopping, and is promoting technological development of construction methods that can be used in sections where persistent river levees are to be built. In this study, a model of a block used in actual river construction was used to measure the hydrodynamic forces exerted by overflowing water, using the shape of its surface protrusions and other parameters. As a result, it was confirmed that when the height of the protrusions on the block surface is increased, the area of drag force action increases, but the drag force decreases because the flow velocity is reduced. The results indicate that constructing open holes in the block reduce the lift force. An example of a highly stable covered block shape for use on the back slope of a river levee s is shown in this study.

 In this study, the behavior of flow around cage system structure, cage deformation and filling stone in the levee overflow was examined experimentally in order to grasp the fundamental performance for the overflow in the case of placing cage system structure in the back slope and the back slope of levee. To begin with, this paper verifies the flow in and out of the cage layer by carrying out the experiment in which cage thickness and particle size of the filling stone were changed and overflow water depth was changed using a large horizontal open channel. Next, the deformation of the cage before and after the overflow and the movement of the cage were measured by 3D scanners. As the result, it became clear that the flow rate became constant from the depth of 4 division of the cage thickness in all cases, and that the flow rate was reduced for the flow rate of the cage surface, regardless of overflow water depth. In addition, this paper confirms that it has sufficient erosion resistance, because the deformation of a cage and drastic movement of a clogged stone are not recognized in the front and rear of overflow.

 In this paper, we focus on the phenomenon of upstream migration of waterfalls in bedrock rivers. We conducted laboratory-scale experiments to investigate the process of upstream migration in mild slope channel, focusing the longitudinal changes in flow velocity and sediment transport velocity that cause this migration. The experimental results showed that step erosion was observed in the supercritical flow section immediately upstream of the edge of the waterfall. PIV and PTV measurements on the fixed bed revealed that the spacing of the vertical sediment transport velocity increase in the longitudinal direction was similar to the step spacing.

 We performed flume experiments to study the effect of the amount of sediment supply on the migration process of a knickpoint formed on bedrock. We supplied sediment in a flume with a modeled bedrock with an initial drop, and observed the migration process of the drop. We find that, if the initial drop is sufficiently large, the migration process of a waterfall is of an overhanging type when the sand supply is large, but of a vertical erosion-dominated type when the sand supply is small. Under the experimental conditions of this study, there is a threshold that separates the migration process at a sediment supply rate between 25-35 g/min.

 It is important to view erosion of valley beds and banks by debris flows not only as an instantaneous and localized process of change, but also as a phenomenon with a spatiotemporal extent. By considering debris flow erosion from an energetic viewpoint, it is possible to consider the erosion mechanism on a spatiotemporally extensive scale. In this study, we investigate the erosion mechanism of valley beds and banks erosion from the viewpoint of the three-dimensional equations of motion, based on detailed three-dimensional numerical experiments on a debris flow. Next, the hydraulic quantities obtained from the three-dimensional equation of motion are used to examine the mechanism of valley beds and banks erosion on a large scale from the viewpoint of energy transport. The results of the study indicate that the valley banks erosion is caused by debris flows with large longitudinal and transverse energy gradients striking the valley banks.

 Rivers with exposed cohesive materials present disaster and environmental concerns. Therefore, it is necessary to understand the phenomenon of gravel deposition and detachment on clay river beds. In this study, we calculated the pick-up rate and adhesive force. The results showed that the larger the bottom friction force is, and the value of the pick-up rate tends to be larger. In addition, the adhesive force shown by the tensile test was within the range of the adhesive force obtained from the force balance in the gravel attached to clay river beds. This suggests that it is possible to obtain the conditions for the detachment of gravel attached to clay river beds from the characteristic values of clay obtained from the tensile test.

 In Japan, gravel extraction and river channel excavation conducted until the 1970s reduced the amount of sediment in river channel and exposed weak bedrock layers. Although various studies have been conducted to contribute to the management of bedrock rivers, the bedrock erosion rate due to collisions of mixed grain size sediments has not been investigated. Therefore, in this study, we conducted an experiment in which mixed grain size sediments were supplied on an erodible mortar bed that simulated weak bedrock, and observed how the sediments were transported and deposited while eroding the bedrock. Our results suggest that not only the mean grain size of mixed grain size sediments, but also the bedrock roughness and the ratio of large and small particles affect the bedrock erosion rate.

 Considering that all sediment originates from the riverbed, this study focuses on the concentration and velocity distributions of bed and suspended load to confirm the continuity of these distributions. Improvements in analytical techniques have made it possible to precisely track and count particles, and not only visualize vertical classification, but also directly calculate the volume of sediment from the concentration distribution. As a result, it was confirmed that suspended load and bed load have similar kinematic characteristics, moving at a high velocity in a downdraft, and the slight relationship of suspended load and bed load classifications was demonstrated once again. In addition, although the two particles were of the same size and no hiding occurred, a clear vertical classification was confirmed, and it became clear to what extent the sand feeding particles contribute to the exchange of the river bed. Furthermore, the experimental values of the concentration distribution and flow velocity allowed us to express the kinetic characteristics of the scouring particles on the riverbed surface in terms of the existence ratio of the particles.

 Present study discusses the sediment sorting process of suspended sediment and bed-load layers based on a new method to describe the suspended sediment transportation. The sediment sorting is inferred to be caused by the coupled phenomena from the entrainment of bed-load layer to the suspended sediment layer and by particle settling from the suspended sediment layer to the bed-load layer, and the mass conservation equation is proposed for each sediment particle size in the flow with suspended sediment as well as in the bed-load layer. Then, we selected the Sen River, which flows into the Tonle Sap Lake, as one of the suspended sediment dominated rivers, and presented survey data on bed material in the channel reach, which confirmed that the sediment sorting of riverbed materials is caused by the proposed process. Besides, it is derived that the grain size distribution of the channel bed surface layer is determined by the setlling process of sediment particles constituting the suspended load.

 A method for classifying land cover of riverine regions into categories such as grassland, woodland, bare land, and water bodies was investigated using a random forest (RF) algorithm. The target sections of the Kamanashi and Kuzuryu rivers were divided into two domains and training and test data sets were generated. The classification accuracies of the RFs were tested in four cases by combining the data sets. Upon applying the method, the highest macro-F1 unweighted mean of F1 scores calculated per class was 73%, and the F1 scores were approximately 90% for water bodies and bare land and approximately 70% for grasslands and woodlands. Furthermore, the accuracy was affected less by the training data sets used to develop the RFs, which may partly be explained by the fact that the digital surface model, vegetation height distribution, and reflected laser intensity are the features that make the highest contributions in the RFs.

 In order to estimate the effects of trees in a river channel on flow capacity and phenomena such as scouring and fixation of low channels using a hydraulic analysis model, it is necessary to determine the density and height of trees in the river channel. Although it takes an enormous amount of labor and cost to survey tree height and density in the entire river channel, it is expected to develop a technology that can efficiently obtain information on trees in the entire river channel by utilizing aerial laser survey (ALB) data, which is widely used in river surveys. In this study, the density of trees (bamboo forests and woody plants) was estimated using ALB point cloud data and an individual-tree analysis method, and the results were used in a hydraulic analysis to evaluate the effect on the simulation result of water levels. Comparison with the field survey results showed that the individual-tree analysis method can extract the height of trees and the range of woody. It was found that the accuracy of the water level simulation result was relatively high when the extracted tree density was used in the simulation model.