A record-breaking heavy rainfall occurred in Hokkaido in 2016, and caused two levee breaches around the confluence of the Satsunai and Tottabetsu rivers, resulting in serious flood damages. We conducted field researches, image analysis of UAV photographs, and flood simulation to verify characteristics of the flood disaster. The flood caused by the levee breach because of main channel migration of the Tottabetsu river at the upstream of the confluence, and inundated landside areas, and subsequently returns to the Satunai river at the confluence by overflowing the levee of the Satunari river. The overflow causes high flow velocity within the inundated landside areas, leading the severe incision of the landside foundation and resulted in another levee breach at the confluence. It means that the high-risk sites of the levee breach would not depend only on the height of the levee in stony bed rivers in which landside could be significant eroded.

The largest recorded flood in the Naka River occurred in August 2014. The water levels in the upstream reach of the elbow section were higher than the design high water level because the water level rise occurred by form resistance of elbow section. It is necessary to develop new countermeasures for reducing of water level rise at the elbow section. We proposed the method to remedy the channel form using the embankment including boulders. We investigated effectuality of this method by using the 2-D riverbed variation analysis and the model experiment. It is shown that this method have the advantage of being able to lowering of water levels in the upstream reach of the elbow section.

The 2D free-surface flow model with the treatment of a bridge was constructed. The model treats a bridge as local energy losses. The validity of the model was shown through verifications against experimental data of the dam-break flow with different bridge geometry. A flood and inundation flow simulation in the Yamakuni River basin at Kyushu-hokubu heavy rain in 2012 was performed with distributed hydrological model as well as the model with a simple treatment of a bridge. The simulated results compared with the observed discharge, water level of flood marker and maximum inundation area. It shows that the model has ability to reproduce runoff, flooding and inundation process with reasonable accuracy.

The numerical simulations of rainfall-runoff and flood flows in the Hikosan River basin were performed and the parameter for the rainfall-runoff simulation was estimated. The rainfall runoffs were simulated by a distributed hydrological model with radar-rainfall. The flood flows in the main river and plural tributaries were simulated simultaneously by the quasi-2D unsteady free-surface flow model. The Genetic Algorithms(GA) and Particle Swarm Optimization(PSO) were used for the optimization method. The parameter optimization was performed against the flooding event in the Hikosan river basin in July 2012 based on the observed water level. It shows that the model and optimization method are a useful tool for examining the process of runoff and flooding and the parameter characteristics in the Hikosan River basin.

This study investigates the effectiveness of the flood risk evaluation method considering spatial distribution of inundation probability by applying to Kinu-river flood disaster in 2015. This method can treat explicitly the uncertainty of location and size of dike breaks. The result shows that the extent of the inundation area of the disaster can be captured by the high probability area on the probability distribution map estimated by this method. Then the influences of floodplain separation by a road embankment on flood risks are evaluated. The floodplain separation reduces the inundation probability and economic damages all over the region though maximum inundation depth increases in some areas.

Several Typhoons attacked the Tokoro river basin in Hokkaido at August, 2016. The maximum rainfall amount for a month of this basin was recorded at this time. At the Futochanae observatory in the downstream part of the Tokoro river, the water level exceeded the design high-water level 4 times during the flood. Overtoppings of the flow happened at several parts around this area. Collapses of the bank arose in the tributaries of the Tokoro river. The river structures in the basin were damaged heavily. In order to grasp the cause of disaster of structures, observations were conducted from immediately after flood damage occurrence. Each cause of disaster is complicated and there was the necessity of synthesizing and examining the knowledge of not only river engineering. Field observation was conducted and the result which contributes to future disaster prevention was obtained from the viewpoint of river engineering, geotechnical engineering and bridge engineering.

Many bridges were broken due to a large channel change during the Hokkaido heavy rain in 2016. We investigated the damage situation of the four bridges (the Kyusen Bridge, the Chiroro Bridge, the Kobayashi Bridge and the Kiyomi Bridge) immediately after the flood. Then we analyzed the collapse mechanism of the Kyusen Bridge using a two-dimensional riverbed deformation model (iRIC – Nays2D).

According to the observation result, the bed height near the Kyusen Bridge rose about 2 m. This implies that sediment supply was large during the flood. The results of numerical simulations indicated that the river bank was eroded largely due to both of a large flow discharge and a sufficient supply of sediment.

In August 2016, heavy rainfall due to Typhoon No.10 caused violent flood damages along the Omoto river. The Omoto river flows through a valley plain, which is composed of a meandering stream channel with sand bars. This study aims to investigate the characteristics of flood flow for obtaining the data on effective river management in mountainous areas based on the results of 2-D numerical model simulations with various sediment supply conditions. The flood flows are compared with various sediment supply conditions in the Otomo-area and Akashika-areas, where severe damages are observed. According to the results of numerical simulations, it is revealed that the water level over the sand bars is obviously affected by the sediment supplies from upstream. Therefore, it is implied that the river planning in the valley plains such as the Omoto river should be based on appropriate estimation of sediment supply from the mountains.

This study aims to investigate flow resistance and its influence on hydrographs during a flood in the Omoto River. Flow resistance at the Akashika station where water level was observed during the target flood was studied, focusing on form resistance due to sand waves, vegetation over floodplain and eddies due to interaction between floodplain and main-channel. The result show that integrated roughness coefficient increases possibly up to 0.061 m^-1/3s at maximum water level with developed sand waves, on the other hand, integrated roughness coefficient takes value of 0.052 m^-1/3s in the transition regime. Computed results on rainfall runoff process considering roughness change discussed above show that especially water level has a high sensitivity to flow resistance, and that simulated water level reproduced observed one accurately in the case that increase of roughness is incorporated into a model.

In August 2016, Hokkaido region of Japan was exposed to record breaking precipitation by four typhoons. The Pekerebetsu river which is a tributary of Tokachi river in Eastern part of Hokkaido, a comparatively small river, was worst affected after facing the last typhoon. The width of the river expanded from two to five times overnight. In this study, the authors conducted several field surveys, and followed by performing numerical calculations to analyze the erosion and deposition phenomena. The results showed that although overall the river width widened significantly, the situation of the channel deformation was different between the upstream and downstream. Numerical analysis explained that the location where sediment accumulated between the upstream and downstream worked to stop the bed rock exposure propagating into the downstream. While, in the downstream, it became difficult to suppress the channel meandering in this disaster scale only by the sediment control facilities in the upstream.

There was a record-breaking rainfall in eastern Hokkaido during the end of August 2016. This rain and associated flood caused the levee breaches at two places and inundation disaster in Sorachi River in Ikutora region, Minami Furano town, Hokkaido. The field surveys and the aerial photographs taken after the flood showed that the inundation disaster was caused by the levee breach at the upstream reach of Ikutora, and this flood flow returned to the Sorachi river main channel from the floodplain, resulting in another levee breach at the downstream reach. Subsequently, we performed numerical simulations of the flood event and associated river bed evolution and levee breaches to investigate detail processes of this flood disaster. The results indicated that the upstream levee breach was caused by both bank erosion and overtopping flow and the downstream levee breach was caused by the overtopping flow.

The frequency and magnitude of the water-related disasters have been increasing all over the world. Several organizations including JSCE flood disaster investigation team generally conduct an on-site survey after severe disasters in order to clarify the damage mechanism. However, there is no solid guideline for the river flood disaster in Japan, whereas the tsunami disaster. This report provides essential information of a proposed guideline for standardization of an on-site field survey and data treatment for flood disaster. We especially focused on ways to identify inundation marks (depth or height) and to evaluate building damage and bank failure. For surveyor’s convenience and safety, the guideline includes concrete cases with many example photos and useful data sheets. The guideline also deals with a simple and fast data sharing method with Google My Map service, a tool for data combination of field data and photos developed on Excel VBA and secure data sharing method provided by JSCE.

A method for estimations of discharge and roughness coefficient at the time of peak water level by floodmarks was examined. The method uses temporal change of water surface profile as input conditions for the estimations. The method also uses 1D shallow water equations applicable to channels of arbitrary cross section as governing equations and flux-difference splitting scheme as a numerical technique. The estimation method was verified against experimental data of unsteady flow in straight channel, gradually contracted channel and expanded channel. The estimation method was also verified against observed data of the flood flows in the Onga River. These verifications show that proposed method can reproduce the discharge as well as roughness coefficient at the time of peak water level by floodmarks with reasonable accuracy.

In this study, we combined the regional map of the location and floor number of the buildings, with the map of the flood depth distribution under the largest flood expected. This combination map was used to analyze factors which would hinder evacuation of the residents and the distribution of the population who would have difficulty evacuating at the time of the disaster because of those factors. From the results, the methods to examine different human damage measures were developed and verified for two regions with different geographical conditions, where there were many residential buildings with difficulty evacuating in a vertical direction and the shortage of public shelters.

We attempt to estimate the maximum flood discharge (extreme discharge) that have occurred within the recent epoch (that is, within approximately the last 10,000 years) based on geomorphological and geological traces of rivers that flow through erosional valley-floor.

We focused on terraced landforms formed through the action of rivers flowing along erosional valley- floor and assumed the terraces along the valley floor to be erosion traces formed during a flooding event. In this paper, we used radioactive carbon dating to date the sediment of flood deposits on the terrace surface , and estimate the extreme river discharge and the process of erosional valley-floor formation. In addition, we verified the validity for discharge per unit drainage of extreme discharge in contradistinction to Creager curve.

We evaluated the risk characteristics prescribed by the probability of various flooding depths and the vertical distribution of asset damage amount depending on building use of housings and offices at their locations. We also evaluated the reduction of asset damage amount with flooding countermeasures such as flood barriers, by giving our vertical distribution models of assets and the probability of various flooding depths. Through these procedures, it is expected economic rationality of flooding countermeasures can be judged for individual building. Furthermore, spatial damage characteristics were compared and discussed through applying the above approach to the 5 districts of cities in Japan, in combination with GIS data which show the location with altitude, floor number, usage etc. of each building, assuming virtual probability of various flooding levels.

We have tried the improvement of searching accuracy of the similar typhoon research system combined the deep neural network, which extended the artificial neural network in the system before the first generation. Three parameters, which are not only the course of typhoon but also the temporal central pressure and speed, were applied to the search engine. It is shown that these parameters were able to improve accuracy to search past typhoons similar to the characteristic of the real typhoon. Furthermore, the function to support the advice of disaster prevention action and flood fighting was estimated by using the results from the search system..

Since sudden localized torrential downpours have been increasing in recent years, it is required to establish disaster prevention / mitigation measures for more speedy and accurate flood control activities and evacuation guidance, etc. We developed an inundation forecast program for processing data on inland and river floods as a whole by integrating the models of river and sewerage hydraulic analysis and flood analysis covering the basins of Kanda River. Incorporating this program, we built a system having the functions to collect the real- time data, convert the data for input in the inundation forecast program, and convert computation results into the data format for distribution to distribute them on the web. This system is expected to support speedy and accurate disaster prevention / mitigation activities in urban areas.

This report presents an investigation for the establishment of river control facility design against tsunami run-up in frozen rivers. The tsunami generated by the 2011 Tohoku Pacific-Coast Earthquake was characterized by the generation of tremendous number of tsunami debris and drifts. In addition, in Hokkaido where ice formed in rivers, broken ice pieces were found in the field survey. The main aim of this report is to propose an optimum river management facility design in the event of winter-season tsunami considering the existence of ice floes. An experimental model of a sluice gate is developed with ice floe models on the water surface and bore-type tsunamis are generated in the flume. The result showed that in some cases the existence of ice can partially attenuate the wave impact load. However, it is confirmed that ice floes also can cause increase in the wave force, which can result in the functional failure of the sluice gate. The experimental results are summarized in the relationships between non- dimensional wave height and maximum wave force.

A retarding basin, that temporarily stores storm water to reduce downstream flow rate, is one of the most effective countermeasures to mitigate existing downstream flooding problems. Effectiveness of the basin is dependent on design of an overflow levee. Experimental works have been commonly employed for the overflow levee design. The objective of this study is to demonstrate that the numerical model, which are comprised of the distributed hydrological model, the 2D unsteady flow model based on flux- difference splitting scheme as well as unstructured finite-volume method, is an effective tool to evaluate the stage-discharge relationship of the Mutabe retarding basin in the Rokkaku River.

We obtained the method to understand the width in evacuation time that includes the uncertainty of the residents' behavior. The uncertainty of starting-point, shelter selection, moving path and movement speed is included in this method. By this method, it can be grasped width in evacuation time structurally as a regional characteristic. Possibilities of the use of this method are indicated as follows; 1) Public administration can understand the allowable boundary of the uncertainty in evacuation. 2) It can be clarified the advance measures for public administration to restrain unallowable uncertain-behavior. 3) Public administration can focus and consider advance measures to shorten the evacuation time.

Kanto-Tohoku heavy rain brought about overflowing and levee breach in Kinugawa River, and about 40km² areas were inundated in Joso-city, Ibaraki prefecture. Thousands of residents were isolated and about 4,300 people were rescued by helicopter, and the delay of evacuation was pointed out. This study evaluated the relationship between residents’ evacuation behavior and a variety of information such as the river related information about river water level and the inundation or the evacuation related information. In order to discuss the methods to use river related risk information for appropriate evacuation, Following contents are conducted. 1) analysis of hearing survey to residents, 2) analysis of inundation expansion by flood simulation, 3) discuss the influence of the disaster information and the inundation situation on residents’ evacuation behavior.

Prevention method of compound channelization using a hydrobarrier spur dike is discussed in the paper. A hydrobarrier spur dike is a permeable spur dike which has some wide slits for the water flow. It is considered that the hydrobarrier spur dikes can erode the deposited sediment in the downstream area and deform the compound channel cross-section. A hydrobarrier spur dike is installed on the bar which is formed in the Azuma River, Aichi Pref. and the effect of the hydrobarrier spur dike is surveyed after the installation. In addition, two-dimensional bed deformation analysis is performed to discuss the suitable location and shape of the hydrobarrier spur dike. The results show that the hydrobarrier spur dike prevented the sedimentation more than the case of the normal spur dike.

Various types of drop structure are constructed to stabilize riverbeds by reducing enormous energy of flood flows and imbalance of transported sediment. Energy dissipating effect was estimated for multi- drop structures of rock-connecting type introduced in Northern anastomotic branch in the Kiso river by using a vertically averaged two-dimensional numerical model implemented in iRIC. The most suitable geometrical condition are studied to install this multi-drop structures to reduce flow energy more effectively by changing discharge and by arranging shapes of drop structure on the basis of 3D-Laser scanning data. The numerical analysis results clarified that changes in water surface profiles, Froude numbers and energy loss quantity depend on locations of hydraulic jump forced by the structures and that flood flows over the structures maintain super-critical conditions when flow discharge exceeds 110m3/s. Also, it is pointed out that multi-drop structures can cause higher energy reduction effect by raising the top heights of each drop structure.

For the temporary coffering method in river works, large sandbags are widely used because of its workability and economical reasons. However, there often can become an impact on construction sites and hinterlands as large sandbags can fall down or be washed away by flooding. Generally, large sandbags are constructed in the coursed masonry rangework form and they are stabilized by their own weight. They can be unsteady because of the fluid force by flooding, as they are not connected nor unified. This article will present the construction case of Continuous Collapsible Wire Mesh Frame with non- woven fabric liner applied the temporary coffering method of torrential rivers. The cases show the result of reinforcement Continuous Collapsible Wire Mesh Frames with non-woven fabric liner against fluid force driven by freshet in Typhoon. They function as cofferdams without damage or being washed away. The article also shows the sustainable flow rate of flooding with the flow regime analysis.

In the event of a disaster in Iwate Prefecture and Hokkaido by Typhoon No.10 in 2016, prevention of disasters in small and medium rivers is considered as an important subject to achieve at present and in the future. Based on these circumstances, in disaster recovery, it is thought that it will become increasingly important to make every possible effort for the infrastructure measures as a prevention of recurrence of disaster. On the other hand, it is observed that in various cases small and medium sized rivers have been suffered by disaster repeatedly in a relatively short period of time. Therefore, the authors implemented the field surveys in three rivers which were selected based on organizing and analyzing the disaster application documents. As a result of the field surveys, the points which need to pay attention in disaster restoration have been found.

At flow attacking points against bank protection works in the steep rivers, local three-dimensional flows and free-surface spatial-temporal fluctuations cause river bed scouring and bank erosion. Three- dimensional flow structures and free-surface fluctuations are important to understand river bed scouring and bank erosion against bank protection works in steep rivers.

In this study, we have two objectives. First, we show structures of three-dimensional flows and free- surface fluctuations against flow attacking points by using the computation results of the General Bottom Velocity Computation method. Second, we show the evaluation method of forces brought by three- dimensional flows around gravel river bank. As a result, forces of three-dimensional flows and free- surface fluctuations are clarified in relation to bank erosions, and the forces on gravel river bank are evaluated by the vector sum of bottom pressure gradient and bottom shear stress.

The most utilized flood runoff model for engineering applications in Japan is based on the storage- discharge function of the type S=kQp. The parameters p and k vary between flood events, causing difficulty in estimating and forecasting the flood runoff hydrographs accurately. The aim of this study is to identify the parameters in the storage-discharge relationship for floods for the Sameura Dam basin, Sagae Dam basin, Taisetsu Dam basin and the Iwaonai Dam basin, located in different regions in Japan and representing different topographical, geological and climatological conditions. In order to optimize the parameters p and k by maximizing the Nash-Sutcliffe efficiency, 10,000 simulations are carried out for each flood event by changing the values of k and p. The results show that the optimum parameters in the storage-discharge function, the sets of p and k, for floods in each study basin are represented by the p- k curve on the log-log graph.

Rainfall-runoff process is the central issue of hydrology. However, the deterministic rainfall-runoff models which are widely used nowadays have their limits. Thus, stochastic method had been suggested. Recently, K. Yoshimi and T. Yamada have tried to use Fokker-Planck equation to study the uncertainty of stream flow due to the random fluctuation in precipitation. The present study is based on K. Yoshimi and T. Yamada’s theory and tries to apply it into practical problems. Especially, the present study suggested a new method of forecasting runoff by using the Fokker-Planck equation in the first step of Kalman filter. Moreover, the present study also generalizes the method to consider the uncertainty in systems which are controlled by partial differential equations.

Conventionally, the riverbed topography has been measured for each cross-section by 200m pitch. In recently years, according to the improvement of measuring technology, two-dimensional surface meas- urement has become possible. The spatial density of measured topography data has become several 10 to 1,000 times higher than conventional measurement. In this paper, we reviewed specific examples of strat- egies and general considerations to utilize surface measurement data of the riverbed topography for the river management.As a specific example, we analyze the result of riverbed topography by swath-type sounder at the downstream of Shonai River. The case study of ditection and monitoring of the defor- mation, grasping the amount of change from the temporal variation of the measurement, grasping the bars and dunes. As a matter of consideration, we complied the spatial resolution and measurement cost.

River channel discharge capacity is not appropriately evaluated by means of the temporally sparse measuring method because the topographical changes in sandbars and growth rate in riparian forests are substantial. Recently, technology (SfM) that determines three-dimensional topography from UAV images has been employed, but this is limited in use for river channel management owing to the lack of topographic data under vegetation cover and the inadequate validation of the precision of measurements. Accordingly, in this study, the verification and evaluation of the measurement accuracy of topographic elevation data (DSM) derived from UAV-SfM for surface conditions (bare, vegetated, or underwater) and vegetation types were performed to develop a river channel management method by easily estimating channel capacity. Furthermore, we devised a method to facilitate the measurement of ground elevation under vegetation cover, which is important for evaluating discharge capacity, and studied the workflow of channel management using topography and imagery data obtained from SfM. Our study indicated that UAV photogrammetry can be effectively used for sensitivity analysis and for confirming ground deformation in low water channels and declining flow capacity due to topographical changes and tree growth.

Aerial photogrammetry of riverbed using UAV and SfM-MVS suffers from a systematic overestimation of the bottom elevation, due to the refraction of light at the water surface. One feasible correction to this effect is to multiply a factor to the apparent depth estimated by the usual procedure.

In this study, we first simulated the geometrically optimal correction factors under various conditions and showed that the widely used correction factor (1.34) is far from being optimal. Then, we proposed a new empirical method to calculate the correction factor appropriate in each flight mission. We proved the superiority of the proposed method through applications to two sites in Saba River, Yamaguchi prefecture.

An orientation measurement method that uses auxiliary recording instruments was developed to enhance applicability and operating efficiency of the image-based non-contact river flow measurement methods such as the Large Scale Particle Image Velocimetry (LSPIV) and the Spatio-Temporal Image Velocimetry (STIV). With this method, field observations can be carried out without the reference marks at the time of observation, consequently saving the efforts of orientation measurement. However, the influence of the angle measurement error on the accuracy of the converted image and the applicable range were not sufficiently studied in the previous research. Therefore, verification was carried out with the objective of presenting proposals focusing on practical aspects such as precision characteristics of the method, practical application range, concept of accuracy management.

It is important to monitor and estimate the change in the water surface profiles along banks over time precisely, to reduce the flood damage. In this paper, we present the assimilation method the water surface profiles by using the adjoint method into the two dimensional shallow water flow model, and applied the method to two floods in 2011 and 2015 in a reach of Tone River between 133km and 152km.

The shape and elevation of the river and vegetation are measured from a helicopter with Airborne LiDAR Bathymetry (ALB) to make a numerical river model. The rate of flow, roughness coefficients, and permeability coefficients of trees are estimated automatically in the adjoint model. In about 10 times of iterations, the validity of data assimilation of water surface profiles is shown in the computation results.

In this study, the evolution of the microtopography on a channel bed of the Shonai River during a flood was studied by using the Structure from Motion-Multi View Stereo (SfM-MVS) technique in order to investigate the influence of riparian herbaceous vegetation on the mechanisms of fine sediment trap. The high-resolution surface models of the Digital Surface Model (DSM) which represented the topography of the pre- and the post-flood periods were produced by the results of the SfM-MVS applications. The numerical calculation by using a two-dimensional bed-evolution model was also implemented to evaluate the influences of the hydraulic characteristics of the reach-scale channel topography on the sediment dynamics. The comparisons of the above results showed that the herbaceous vegetation tends to hold fine sediments during the flood. The micro-highlands which were grown by the trapped fine sediment by the vegetation tend to remain on the same places as well, even though they located on the regions where the numerical calculation results implied which tended to be eroded.

By applying handheld X-ray fluorescence (hXRF)―one of the most affordable and rapid elemental analysis methods―to various sizes of riverbed sediments, the effect of sediment diameter variation on hXRF analysis was investigated. As a result, large-diameter sediments had large variation in elemental composition, and some element abundances varied with grain diameter. Therefore, it is necessary to consider the diameter of riverbed sediments. Next, hXRF analysis was applied to estimate the sediment distribution pattern in the Mimikawa River, Japan. In a three-element (Rb–Sr–MnO) diagram, large-diameter sediments showed larger differences in element abundances down the length of the river as compared to small-diameter sediments. Hence, the use of elemental compositions analyzed by hXRF is useful in efficiently estimating the sediment distribution pattern of each diameter of sediment.

This study aims to show the validity of phosphorus reduction of oyster shell, which is installed at a moving hydrofoil drone, as countermeasures of eutrophication in rivers and ponds. For this, we conducted laboratory of oyster shell on phosphorus reduction and field tests in a small urban river and small pond. The results in laboratory experiments indicated that the oyster shell treated at 800 degree with size of 2–4.75 mm was a high function of phosphate removal. It is also found in the laboratory experiments that the moving-type oyster shell has a higher function of phosphorus reduction than fixed-type ones. The phosphorus reduction due to moving hydrofoil drone with oyster shell is evaluated based on the laboratory and field experiments, showing the high performance of the moving drone with oyster shell.

This paper explores the possibility of utilizing the video feeds from the existing river surveillance cameras in Hyogo Prefecture for the purpose of estimating the flood flow rate by STIV(SpaceTime Image Velocimetry) method. Utilizing the IP and analogue camera video feeds at ten-minute interval for the STIV analysis, it was found that the estimated results achieved the correlation coefficient of around 0.8, and was deemed capable of producing a Height- Quantity Curve with an acceptable level of accuracy. It was noted that the image resolution from the analogue video feeds at 30 fps was higher than that by the IP camera video feeds at 5 fps. While the night-time STIV analysis is challenging, at spots where the surface ripple pattern were visible by the adjacent roadway illumination and other light sources, usable data could be obtained for the purpose. It is concluded that, by archiving video feeds taken during the flood events and improving the suitability of such feeds for STIV analysis, a sustainable implementation framework for river flow rate monitoring can be established at numerous locations within the prefecture.

In this study, the authors developed an automatic discharge measurement using ADCP and DIEX method. We observed the four different hydraulic phenomena downstream of Teshio river. 1) flow at the time of normal and flood stage, 2) flow at the salinity intrusion, 3) flow at the covered with ice, 4) flow at the salinity intrusion while being covered with ice. In the case of the normal and flood flow, the discharge was calculated by DIEX method using horizontal flow velocity distribution of H-ADCP. When the salinity intrusion or being covered with ice, firstly, we estimated the fresh-salt water interface and bottom height of ice using the velocity profile and bottom tracking of V-ADCP. Secondly, we calculated the unit width discharge to measured velocity profile. Thirdly, we calculated the logarithmic law flow velocity profile of the open channel flow with same the unit width discharge. Finally, the discharge was calculated with DIEX method using one representative point of the logarithmic law flow velocity profile. The discharge accuracy was about ±10% under the four different of hydraulic phenomena. This result showed that it was possible to calculate the discharge automatically without man working.

Small and medium rivers have characteristics that are susceptible to localized and extreme precipitation. Therefore, the hydrologic data is important for river management of small and medium rivers. However, there are some problems which the shortage of financial resources for river management and chronic shortage of observer at the flood condition. In this study, the authors developed an unmanned river flow velocity measurement system using the space-time image velocimetry(STIV) on small and medium rivers using a low cost trail camera used for monitoring animals and plants. In addition, this camera could be taken a shot all day because it had the function that near-infrared image. Measurement results in four different rivers were within ± 10% error compared with the approval flow meter. This method is an effective method for river management of small and medium rivers.

While short-time forecasting using XRAIN data has been developed in recent years, mesoscale weather predictions with XRAIN data assimilation have been limited. This study discusses how and to which extent the prediction accuracy of WRF (Weather Research Forecasting model) can be enhanced with XRAIN data assimilation for the torrential rainfall in Hiroshima city. The predicted results show that data assimilation leads to significant improvement in the location of linear rainbands, since the data assimilation successfully induces the favorable environmental conditions for convective rainfall. It is clearly shown that the analysis cycle of 10 minutes produces better results than those at 30 minute interval and an hour interval. It is also clarified that the assimilation of XRAIN data at only lower angles does not improve the prediction accuracy because cumulonimbus clouds developing in upper layer can be well captured with the radar data at higher angles.

Long-term trends of yearly maximum water level (YMWL) in 50 Japanese Class A river systems have been investigated by applying t test (parametric) and Mann-Kendall test (non-parametric) using hourly water level observed at 471 stations during 30 years from 1986 to 2015. No significant trend of mean YMWL in 7 regions and Japan were found with 95% confidence level, while YMWL slightly increased in most of regions and Japan. On the other hand, YMWL at 31 stations significantly increased with 95% confidence level, and that at 13 stations significantly decreased. The slope of the trend of all stations was ranged between -0.1 m/yr and 0.1 m/yr, and was shifted to positive domain. Therefore, our results indicate that the flood risk can increase during 30 years in Japan.

The paper discusses the necessity of the integrated river plan to raise the safety level against water-related disasters in the near future. It is investigated that how water allocations in every reach in rivers and in dams were made and what flood control facilities seen over an aerial view of basins were endowed with insufficient functions. The flood water volume distributions in upstream, middle and downstream reaches were investigated from the use of widespread water surface profiles by the numerical model. The new integrated river plan is proposed on the basis of the results of numerical computations using temporary flood water surface profiles in the 2015 Kinu river flood, in which the ship-bottom shaped cross-sections with an elevated river bed levels are adopted.

The Hii River floodway has important role to reduce flood damages in the downstream reach of the Hii River. The flood discharges diverted to the Hii River floodway are controlled by the movable weirs installed at the bifurcation section. The movable weirs are operated based on the rating curve in the observation station. Therefore, estimating flood water surface profiles and discharge distributions at the real-time during a flood are needed for proper operations of the movable weirs. Recently, the multipoint water level observation system which was able to obtain multipoint water levels at real-time was developed. In this study, the 2D flow and bed variation calculation method which was able to estimate water surface profiles and discharge distributions at real-time is presented by using multipoint observed water levels in the Hii River.

This study is to present a new safe and stable monitoring system for river water level and discharge by cooperation of image processing and hydraulic analysis. The authors individually developed subsystems such as (1) CCTV camera control, (2) water surface detection algorithm, (3) STIV and (4) DIEX method, and these were integrated. To confirm the fundamental performance of the present monitoring system, the prototype system has installed into the Ongagawa River, and laboratory experiments have also conducted. The results indicated that the evaluated velocity and discharge give good agreement with the observed data. Furthermore, we proposed a method for rejecting outlier.

In water resource management and disaster prevention in the snowmelt season, it is very important to estimate snow distribution in the mountainous regions as accurately as possible before the snow starts to melt. This paper presents a simple method to estimate snow distribution with high accuracy based on an observed value at one representative site.

We applied the method to the Jozankei dam basin in Hokkaido, northernmost island of Japan, to estimate snow distribution. As a result, total snow water equivalents in the dam basin in 20 years and dam inflow in the 16 snowmelt season were estimated with high accuracy. We can probably estimate snow distribution in the Jozankei dam basin with high accuracy without snow surveys.

For real-time prediction of river water level, a numerical method of adjoint sensitivity analysis in river flow simulation was studied. First, numerical method of adjoint sensitivity analysis to assimilate observed water level into one-dimensional river flow model is presented. Next, the model verification was carried out to assimilate water level and discharge by using hydraulic model tests and identical twin experiment. As a result, this method provided a practically sufficient assimilated value of upstream boundary discharge and water level time series. Finally, this method has been applied to the actual flood event of the Tama River. As the prediction accuracy of the water level and discharge is sufficient, the validity of this method was confirmed.

Once flooding occurs, the river administrator and/or flood protection organization needs to take numerous complex disaster management actions. In addition, more often than not, this has to be done with limited time and staff. This is because Fuji River has characteristics such as fast flood run-off, many flood prone areas, conspicuous river-bed variations, etc. This paper reports on the development of a simplified water gauge and disaster management work support system, which considers these characteristics of Fuji River. Further, it discusses efficiency and labor savings to the flood monitoring system based on utilization in flood correspondence training and trial operations.

In flood prediction, reduction of the uncertainty is one of the biggest issues. As a prediction model, artificial neural network (ANN) is one of the potent methods. To improve the accuracy and lead-time of the river-stage prediction model, it is desirable to take advantage of as many observation data as possible. However, due to the limitation of the learning ability of ANN model, it is undesirable to use too many input data into the model. As a new learning method of ANN, deep learning is said to have an excellent ability of learning, and able to handle huge input data. In this study, we developed to models, 4 layered deep learning based ANN and 3 layered conventional ANN. We applied 2 models to the 3 rivers in Japan, Ooyodo River, Onga River and Kokai River. Compared to the conventional ANN, deep learning based ANN model could reflect more observational information and could reproduce the past flood well.