Marine plastic wastes have been mostly originated from inland, and it is important to monitor macroplastics inflow into oceans via rivers. This study aims to develop a new image processing to capture the area and types of macroplastics with deep learning. The learning data for floating debris were collected with the field test under normal and flooding conditions. CNN and YOLO were applied to find the area and types of macroplastics. The results indicated that the CNN and YOLO can capture acceptably the area and types of macroplastics in normal flow condition. It is noted that add of the learning data under flooding conditions can greatly improve the accuracy of distinguishing the types of macroplastics by YOLO.

 In this study, we conducted the river geomorphology monitoring based on the UAV photogrammetry with the water surface correction in the Saba River, Yamaguchi Prefecture. Moreover, we compared the survey results of the UAV photogrammetry (UAV with RTK and one without RTK) with those of ALB. As a result, by using UAV photogrammetry for monitoring in the river channel, we could grasp the changes in the sandbars, riverbed, riverbed fluctuation amount, and sediment balance for a short span of time. In addition, the survey results of UAV photogrammetry were not significantly different from those of ALB, suggesting that it can also be applied to the underwater surveys. It was also suggested that by using the UAV with RTK, this method can significantly reduce the survey time, keeping the same accuracy as the survey with the UAV without RTK.

 In recent years, river disasters because of heavy rainfall have occurred especially in river confluences, narrowed river channels, and vegetated river sections. Therefore, this situation necessitates rapid countermeasures by probing the current river flow capacity based on detailed data for both river bed elevation and land cover conditions in local river areas vulnerable to the disasters. For this study, we investigated a deep learning method for the land cover mapping in rivers using both the laser point cloud data and digital photographs obtained by a UAV-borne green LiDAR system. Our targeted site was a 1.2-km long government-controlled section of the lower reach of the Asahi River in Okayama Prefecture. Results showed that the deep learning scheme trained using the photographs produces the land cover mapping more accurately, compared with that trained using the laser data. Results also demonstrated that the use of the laser data will rather contribute to improve the quality of the training data.

 To understand transportation mechanisms of materials through an ocean surface by wind, a preliminary study is conducted on a spaciotemporal wind pressure measurement on a continuous solid wavy bed by means of the pressure sensitive paint (PSP). The PSP measurement technology is close to an application stage in the field of aero-fluid-dynamics. However, the wind speed acting on the ocean surface and civil engineering structures is not high enough to apply the PSP measurement. For example, the pressure drop is only 0.54 % for a wind speed of 30 m/s. Therefore, temporal and spatial data accumulation methods are applied, such as binning of intensity data from neighboring pixels, and accumulation of data for repetitive captures of the images. The PSP luminescence intensity is sensitive to temperature changes caused by wind and illumination to the PSP. Therefore, it is necessary to accurately measure the temperature change at the same time to correct the pressure estimated from the intensity or the decay constant of the PSP luminescence.

 The basic principle of the Globe Anemo-radiometer (GAR) is based on the thermal analysis of globe thermometers. Using the characteristic of being able to select three of the four meteorological variables (long and shortwave radiation, wind speed, and temperature), we verified whether it is possible to inverse identification of temperature by measuring longwave radiation separately. When the temperature of the thermocouple with forced ventilation was taken as the true value and the longwave radiation measured by the radiometer was taken as the input value, the RMSE was 0.535 °C, suggesting that the inverse identification of temperature from GAR was possible.

 The double grid model employs two scaled grid systems: a fine grid for storing topographic and roughness data and a coarse grid for solving governing equations in order to simulate large scale flow phenomena efficiently. To apply the double grid model to flood in an urban area, I refined the double grid model by considering the drag force due to the buildings inside the coarse grid cells and strict mass conservation at wet and dry cell transitions. The modified model was applied to the previous flood experiment in Sapporo city area. The computational results showed that the present model can simulate reasonably a flood inundation at an urban area.

 A reduced-order model for flood simulation was developed by using the proper orthogonal decomposition (POD) and the discrete empirical interpolation method (DEIM). The primitive reduced-order equations were first derived by substiting linear combinations of the POD basis vectors for the state variables into the original model equations. The DEIM was then employed to approximate the nonlinear terms of the primitive equations efficiently. The reduced-order model was applied to an imaginary computation domain with several conditions. The numerical experiments revealed that (1) the reduced-order model was capable of reproducing the results of the original model, while the simulation errors highly depended on both simulation conditions and the degree of model reduction, and (2) the computational cost of the reduced-order model was less than half that of the original model for the given conditions.

 When a supercritical flow comes into a bend accompanied with a oblique hydraulic jump, the flow becomes 2D. Ippen’s theory well describes such flows, and we showed that the CRD numerical scheme well reproduce such flows. If both banks have such bend with contraction of the flume, and the two jumps cross each other, the flow becomes more complicated. 2D CRD schemes is applied to the analyses of such flows.

 The present paper investigated the development of a hydraulic bore in such a flow by use of 1D and 2D CRD schemes. The superposition of the two jumps causes a rise of the water level, and makes the Froude number F lower, and when Fr<1, it becomes a specie of a jump. If it extends over the whole width of the flume, it becomes a bore advancing upstream. In addition, the jump water and the wave are affected by the refractive angle. The larger the refractive angle, the stronger the two-dimensional effect.

 Theoretical estimation of discharge coefficient of submerged flow through sluice gate is generally agreed to be difficult, when open height is big and water level of the downstream side of gate is close to water level of the upstream side. Numerical simulation of submerged flow was conducted on the conditions that open height is big, including the above-mentioned case. The results confirmed that characteristics of discharge coefficient estimated by simulation was generally consistent with theoretical characteristics, and that the relations of open height and discharge coefficient were estimated using cubic curve approximation passing through the origin.

 Multi-resolution technique has been implemented to the SPH method to reduce high computational costs in standard SPH with uniform particle size. The multi-resolution model splits a coarse particle into two fine particles in a high-resolution area and merges adjacent two fine particles into a coarse particle outside of the area. To check the validitity and efficiency of the present model, a dam-break and wave overtopping problems were analyzed. The model reproduced the measured location of dam-break tip with reasonable accuracy. The model also agreed well with the high-resolution uniform model both in the dam-break and overtopping problems, however, slightly increasing pressure oscillation. In terms of efficiency, the computational cost was reduced by 40% compared to the high-resolution uniform model with the present configuration.

 Hydrodynamic modeling in reservoirs usually implements the Uniform Distribution Method (UDM) for outflows. While this may work well in single-point portals like for deep penstock withdrawal (DPW), the case may be different for selective withdrawal (SW). Using actual observations of velocity fields through the SW facility of the Ogouchi Reservoir, a new outflow method called the Modified Gaussian Distribution Method (MGDM) was applied to simulate the in-reservoir and outflow temperatures. Results showed that MGDM can reproduce the two thermal properties more reasonably than UDM. Using these techniques, the three conceptual cases were formulated namely, Case A (DPW), B (SW) and C [SW and vertical curtains (VC)]. Case A has wider thermal dispersion and exhibits cold water pollution while Cases B and C develop stronger thermoclines and can mitigate the thermal pollution. Case C has wider epilimnion than B due to the VC yet simulation shows similar trends for outflow temperatures for B and C due to similar release operations. In modeling reservoirs with SW, it is highly recommended to measure the actual velocity field of the outflow and apply these measured velocity distributions in the simulation. MGDM is promising for modeling with better accuracy not only for the reservoir temperatures but also for the sediment and other water quality distributions.

 The wind stress acting on the water surface and the thermal stratification are typical factors changing turbulent flows and material transfers in natural water-environments. In this study, we analyze numerically how these two factors affect turbulence properties and the heat transport flux at the water surface in a turbulent open-channel flow by means of the direct numerical simulation (DNS). Under the stable, neutral and unstable stratification conditions, the open-channel flow with the surface shear stress is simulated, and the surface divergence and the heat transport flux are obtained from the numerical data. The numerical results demonstrate that the surface divergence can be universally described with the Taylor micro scale even in the presence of the combined effects of the two factors. The standard surface divergence model is confirmed to be approximately applied for the turbulent open-channel flow driven by the surface shear stress and the thermal stratification, but the proportional coefficient of the divergence model varies complicatedly with the surface shear stress, depending on the Richardson number.

 The overgrowth of Egeria densa, invasive aquatic plant, is a major issue in the upstream of the Gonokawa River. It is necessary to develop the efficient method for eradicating the nonnative aquatic plant. In this study, we quantified the coverage of E. densa with UAV on the three seasons (dormant, post-flood, and senescence seasons) and analyzed the relationships between the coverage of E. densa inhibiting throughout the seasons (source patches) and environmental factors using generalized linear models. Our results indicated that the source patches tended to exist in the upstream of the survey segment. In addition, the presence of source patches had the highest negative correlation with the fluctuations of the river discharge, suggesting that the river discharge fluctuations could suppress the overgrowth of E. densa in the river segment by preventing the formation of the source patches.

 We aimed to evaluate the effects of sediment replenishment on the riverine environment in the downstream reaches of the Hamahara Dam, Gonokawa River. We estimated the amount of the sediment transported from the replenished sediment by photogrammetry and morphodynamic simulation. We also examined differences in the riverbed environment and fish and benthic organism abundances between fall and winter after the summer floods in 2020. The 30% (2,625 m³) of the sediment replenished in the previous year was estimated to be transported to the site 3 km downstream from the replenished sediment. In the fall, at the site 1 km downstream of the replenished sediment, the values of riverbed penetration and environmental DNA (eDNA) concentrations for sand-dwelling fishes were higher than those of the upstream site. These results suggest that the riverbed environment at the site below the replenished sediment could have been improved by the sediment replenishment. In contrast, the values of riverbed penetration and the eDNA concentrations for sand-dwelling fishes had decreased and the biomass of net spinning caddisfly larvae had increased at the site below the replenished sediment by winter. These results indicated that the riverbed improvement was sustained only for ca. three months after floods.

 It is expected that not only air temperature but also watershed characteristics could affect the river water temperature increase rate. In this study, we examined the relationships between the river water temperature and watershed characteristics in the 10 first-class rivers in the Chugoku District and calculated the increase rate of the water temperature relative to the air temperature increase. We also examined how land use ratio in the watersheds could affect the river temperature increase rate in the study sites. Our results indicated that the river water increase rate could significantly increase as farmland and paddy filed ratios and building ratio increased and forest ratio decreased in our study system. These results suggest that the land development could strongly affect the river water increase rate in the study region.

 Under the circumstance of climate change, it is essential to forecast the future temperature of rivers to protect the ecosystem and other environment. This study applied a model of river channel temperature model coupled with rainfall runoff inundation model to simulate the discharge and water temperature of the Yamato River. Despite of some weaknesses, the Nash-Sutcliffe coefficient is above 0.8 and the RMSE of river water temperature is around 1.85 °C, which means that the calculations of two models were both in good agreement with the observations. Then, we evaluated the effect of air temperature rise, as well as decrease in sewage effluent, which is one aspect of depopulation, on the river temperature of Yamato River. As a result, it is shown that the rise of air temperature poses a significant impact on river water temperature, and the effect of decrease in sewage effluent may not offset the effect of air temperature rise.

 To clarify the factors that govern the water temperature characteristics in a basin, in this study, the observed water temperature and the equilibrium water temperature were used to build regression tree models with the modified thermal sensitivity (dM_mT_w/dM_mT_eq) or the thermal equilibrium difference (M_aT_w−M_aT_eq) as the objective variable. Additionally, 33 characteristics related to the basin and channel were used as the explanatory variables. The variable importance analysis was performed using the developed models. The results revealed that the channel characteristics such as the channel length and slope, and basin characteristics such as the catchment area and elevation are the most important variables for the gradient. Additionally, with regard to the deviation, the variables related to the land use or cover in the catchment areas, such as forest, grassland, and agricultural land, and the aspect-slope that receives solar radiation are the main factors that drive an increase in water temperature.

 Plastic pollution is a global environmental problem, and it is necessary to understand the fluxes of macroplastics (> 5 mm) from rivers to the sea. We developed a monitoring system for quantifying anthropogenic debris flux by applying an image analysis into video images captured by a IP camera system. The image analysis made the color differences among water surface, natural debris and anthropogenic debris clear. We have conducted continuous measurements of anthropogenic debris flux in the Tenpaku River and its drainage channel during 34 flood events. Anthropogenic debris flux in rising stages significantly increased comparing with those in falling stages. The mean concentration of anthropogenic debris in the rising stages was significantly related to antecedent precipitation index (API).

 This study investigates the relationship between distribution characteristics of fluvial morphology and geomorphic characteristics of the headstream area. The gradient of the fluvial morphology at the reach scale is large in the order of fall-pool, cascade, step-pool and pool-riffle. Slope relief and drainage density were selected as mountainous characteristic quantities that affect the longitudinal profile and the distribution characteristics of the fluvial morphology. These indices are considered to be important factors because they are related to sediment production and specific height difference of the stream bed.

 In the Muka River, the Tokoro River system, in 2016.8, the flood caused outflow gravel bed of surface and the exposure of the volcanic ash layer. Sudden degradation of riverbed occurs, and lowering of the stability of the low-water revetment and the bridge with that. In this study, we considered survey method for understanding the gravel bed of surface at the bedrock river. As a result, it was found that the combination of surface wave survey and ground-penetrating radar survey can be a method to determine the thickness of sand and gravel layers more accurately by interpolating the results of both surveys.In addition, it clarified the problem that extend the range of applicability of geophysical prospecting at river area.

 This study is focused on the effect of the surface shape of the dry masonry revetment on the backfilling sediment suction phenomenon, and hydraulic experiment is conducted to verify the effect.

 As a result, it was found that when there is a protrusion on the surface of the revetment, remarkable backfilling sediment suction phenomenon was observed. It was confirmed that the larger the protrusion on the surface of the revetment makes the greater the amount of sediment sucked out. Furthermore, by measuring the pressure inside the revetment, it was found that the pressure near the revetment increased at the upstream point of the revetment and decreased at the downstream point. It was confirmed that the larger the protrusion on the surface of the revetment makes the larger the pressure difference between the upstream and downstream points of the revetment. It was shown that the pressure difference is one of the causes of the suction phenomenon.

 In this study, a simplified sediment deposition rate model was proposed based on field observations of suspended sediment in vegetated areas of real rivers. As a field experiment, sediment trapping panels using simulated vegetation were installed in a small river to verify the physical properties of the deposited sediment. The results showed that the sediment deposited in the vegetated area was basically transported in the form of suspended sediment in the absence of renewal of riverbed materials by the bed load transport. The results by the presented model were of the same order of magnitude as the measured deposition rate.

 In recent years, flood damage has frequently occurred due to the increase in memorial heavy rainfall. At a large food event, resistance by vegetation is greatly changed by trapping floating debris that is produced through the washout phenomenon in upstream. This study classified the trap pattern into three, and examind the resistance characteristics by modeling a trapped material and a forest using real tree branches. I-type trap that debris is only trapped in front of forest can increase the water depth locally. U-type trap that each tree in the forest traps debris can increase the whole water depth and the influence is large. Doubling the thickness of the trap object, the drag coefficient increases around 1.2 to 2.1 times for the I type and 1.3 to 1.6 times for the U type. This is because the flow velocity in the trap object is assumed constant, however it is not constant and decreases while passing, and the resistance from rear side of the object decreases because of the shielding effect.

 In this study, we evaluated the distribution and habitat characteristics of Alternanthera philoxeroides, a specific alien organism, and the vegetation after excavation of high-water channels in the Kikuchi River. As a result, it was clarified that this species is distributed in almost the entire area of segment 2-1 proposed by Koichi Yamamoto, and that it easily invades and settles in the gentle waterfront area where the shear velocity is relatively low, the median riverbank slope is 5%. It was also shown that excavation of high-water channels is one of the factors for the invasion and expansion of this species, and in particular, excavation with a gentle slope at the water's edge has a risk of promoting the invasion of the species. Furthermore, it is shown that many stems may be bundled to cover the water's edge area, which may cause the species to settle and expand in rivers.

 Water lettuce, Pistia stratiotes L. is widespread in tropical and sub-tropical regions. The excessive growth of water lettuce causes ecological and socio-economic problems, however, there is few filed data on water quality under the dense mats formed by the invasive floating macrophyte in Japan. We conducted the field observation in the Turuda-Dam reservoir whose water surface were covered by water lettuce mainly. The results showed that the mats lowerd light intensity, phytoplankton productivity and dissolved oxygen concentrations near the water surface. In addition, it was confirmed that the water temperature was adequate for overwintering of water lettuce in the winter of 2020.

 The influence of the storm events on carbon (C) fluxes is concerned due to climate change. Noticeably, the intense typhoon events may mix and renew the water column in subtropical lakes quickly. Thus, the vertical mixing and residence time are critical physical factors controlling the vertical distribution of C within shallow subtropical lakes. We choose a small, stratified subtropical lake (Yuan-Yang Lake, YYL) to consider thermal stratification and calculate residence time during a strong typhoon by a three-dimensions numerical model, Fantom. The results showed that the CO₂ was released approximately 400-1000 mg CO₂ m^-2d^-1 across the water surface to the atmosphere during the typhoon period due to C loading via sediment and river flows. The typhoon-induced mixing and substantial amount of river flow rapidly intrude into the thermocline, resulting in a residence time of fewer than 2.0 days during the strong typhoon. This study suggests that considering the vertical distribution of water temperature and C helps calculate the residence time and C fluxes in the lake ecosystem.

 A three-dimensional water quality model with an unstructured grid in the horizontal direction and an LSC² coordinate system in the vertical direction was used to perform numerical simulations of the overturn of water and the recovery process of bottom dissolved oxygen (DO) in Lake Biwa. The numerical model reproduced well the seasonal variation of vertical profiles of water temperature and DO in the northern lake of Lake Biwa, and the relationship between water temperature and DO in the bottom layer at Imazuoki-chuo station showed a similar trend to the results of field observations. In January 2015, after the overturn of water with vertical uniformity of water temperature occurred, the lake water with low water temperature and high DO cooled in the shallows along the eastern shore dived into the bottom layer as a density current and entered the bottom layer at Imazuoki-chuo station, resulting in the recovery of bottom DO. Although the timing of the overturn of water depended on the air temperature conditions of the year, the recovery process of bottom DO due to the intrusion of cold water was considered to occur every year.

 Aegagropila linnaei, Marimo, is a special natural monument of Japan and in danger of extinction. However, the formation mechanism of Marimo has been unresolved. In particular, it is needed to estimate the rotation angle of Marimo, which plays a significant role in the formation of Marimo. Numerical computation is one of the most robust techniques. Thus, we attempted to develop the Marimo model, including the Marimo's movement and its rotation. The Marimo model showed a relatively good reproducibility with the laboratory experiments in terms of rotation when the Marimo was in contact. However, when they were not in contact, the model did not reproduce well. It was also suggested that the rotational moment needs to be taken into account.

 In recent years, the number of combined disasters caused by floods and driftwood has been increasing, and countermeasures against driftwood disasters are required. Dam reservoirs have the role of trapping driftwood with a boom, which contributes to reducing flood damage downstream. Although booms play such an important role, the age of booms is unclear, and the timing of their replacement can be determined by appearance. In view of the increasing external force of floods and the amount of inflowing wood, it is important to clarify the deterioration mechanism of booms and to maintain the booms to prevent it from breaking during floods. In this study, we proposed a boom detection method using image processing that is robust against noise and obstacles, in order to observe the behavior of booms during normal and flood conditions from time-lapse images of the booms taken at the Amagase Dam, and to clarify the relationship with hydraulic and meteorological observations. In addition, the correlation between the behavior of the detected net field and the observed wind direction, wind speed and flow rate is investigated.

 Oxygen-deficient water masses occurring in the lower upstream layers of rubber weirs and flows of water with poor oxygen due to the collapsing of rubber weirs considerably affect the water quality in downstream dam reservoirs. Therefore, in this study, we organize and analyze the results of existing water quality survey, elucidate the characteristics under which oxygen-deficient water masses occur, and construct a quantitative model of dissolved oxygen consumption because of sediment. We confirmed the validity of the numerical analysis model through a comparison with actual results. The results of this investigation revealed that oxygen-deficient water masses were distributed in areas with an approximate depth of 4 m from the upstream of the rubber weir to the intermediate region. In addition, The daily minimum total flow volume applied to the collapsing, as calculated using the sediment oxygen consumption rate estimated values, should be at least 0.3~0.8 times that of the reservoir capacity.

 Shallow lakes are more susceptible to the effects of climate and climatic changes than deep lakes and dams because of their small depth. In shallow lakes, diurnal stratification occurs during daytime with mixing during nighttime, although it is not easy to model and analyze the effects of diurnal stratification. In this study, after preparing wind condition and estimation of temperature of inflowing river water, we aimed to improve the simulation of water temperature stratification in Lake Suwa by using a bottom heat transfer model that can take into account the cooling effect of mud at the bottom of the lake. The results suggested that the mud layer at the bottom of Lake Suwa cools the bottom water mass directly in summer period when the temperature is vertically stratified, and the hydrodynamic model coupled with the bottom heat transfer model would effectively simulate the structure of the theramal stratification in summer. Meanwhile, lowering the inflow river temperature as a boundary condition did not have a significant effect on the reproduction of low temperature in summer bottom layer.

 In Lakes Shinji and Nakaumi, which are connected brackish water lakes, there is a growing interest in future changes in the water environment. Therefore, we evaluated the water environment of both lakes for the present (1996–2005) and the future (2090–2099). For the future climate, three GCMs were used, and bias correction was applied. The runoff model of HSPF was used to predict the basin's inflow and water temperature. The water quality of the two lakes was predicted using a hydro-environmental model of AEM3D with these boundary conditions. As a result, precipitation is predicted to decrease by 10–20%, air temperature is predicted to increase by 1.3–4.0°C, and the inflow to the lakes is predicted to decrease by about 30% in the future. Moreover, the water temperatures in both lakes were predicted to increase by 0.7–2.5°C in the surface layers and by 0.5–2.5°C in the bottom layers. For salinity in the future, approximately 1.8 times greater increase in Lake Shinji and the more significant increase in the surface layer than in the bottom layer in Lake Nakaumi were predicted.

 Our study aimed to develop a method to efficiently classify bottom materials in shallow waters. Bottom material pictures taken at Lake Shinji, Shimane Prefecture were differentiated by the UAV photogrammetry and image analysis using deep learning. As a result, the machine learning identified images of mudstones, stones, and sand obtained from orthoimages with the identification rate of about 0.65. In addition, there was no significant change in the identification rate between the case using the underwater image and the case using the orthoimages. Moreover, when bottom material for entire survey area was predicted with this method, the predicted results were similar to the results obtained in the field surveys. These results suggested that this method can grasp approximately bottom material distributions in shallow waters efficiently.

 To clarify climate change impacts on nutrient behavior in Harima Nada located on eastern part of Seto Inland Sea, we performed flow analyses of water and nitrogen based on numerical simulations under the present (the end of the 20th century) and the RCP8.5 future (the end of the 21st century) climate conditions using an integrated river basin – coastal ocean hydroenvironment assessment model. There was no significant difference in water flow between the present and the future climates. However, horizontal flow of nitrogen during summer and autumn under the future climate became higher than that under the present climate because sea temperarute warming reduced primary production and subsequent sedimentation. During winter and spring, seasonal decreasing in nitrogen amount in the future Harima Nada was enhanced by higher primary production because of sea temperarute warming.

 To investigate the effect of the downstream river topography on water distribution at the confluence of Arakawa River estuary, a 3D flow simulator was applied. The simulator was confirmed to reproduce measured data of water level, salinity, and cross-sectional velocity at the confluence during spring tide. The results showed that the flow ratio of the Arakawa and Sumida Rivers was not proportional to the cross-sectional area ratio, which was affected by the water level difference before and after the confluence. Specifically, during low tide, the inflow from Shingashi River suppressed the diversion from Arakawa River to the Sumida River. In addition, the rapid contraction of the river width and the meandering shape of the river downstream raise the water level, which causes the distribution of water level differences at the confluence.

 A numerical investigation has been carried out to quantify the variations of salinity and dissolved oxygen (DO) along the urban river estuary to understand the relationship between saline intrusion and the generation of scum and odor. The numerical model was set up from the middle of the Tokyo Bay to the Shakujii River Estuary (SRE) to take account all hydrodynamic processes from sea to the upstream. The observed and numerical results showed that the type of the mixing along the estuary was changed from well-mixed type to saline-wedge type in the transition from spring to neap tides, and there was no substantial velocity during the spring tide in the upstream of the SRE. The numerical results also showed that DO concentration of intruding anoxic sea water from the Tokyo Bay decreased down to nearly zero before reaching the upstream, indicating significant effects of the saline intrusion on DO environment, i.e., the scum generation in the SRE. Lastly, the numerical experiments revealed that DO concentration of anoxic sea water increased to 4 mg/L in the SRE without the bottom consumption, but decreased to 1 mg/L with it, therefore suggested the importance of the bottom consumption on DO environment in the SRE.

 The 2011 Great East Japan Earthquake and Tsunami caused large-scale damages in many areas. In the estuarine regions, many rivers experienced the flushing of sand spits. In the Natori River of Miyagi Prefecture, the recovery process of estuarine morphology has been observed since the tsunami, but it has not been evaluated whether the morphology has already become stable or not. Therefore, in addition to the analysis of collected aerial photographs, a detailed analysis of the beach topography in front of the river mouth was conducted. By assuming an equilibrium beach cross-section, the approach to this cross-section after the tsunami was confirmed. The convergence process to the stable beach was quantitatively evaluated by the time function obtained from the empirical orthogonal function (EOF) analysis. From the results, the disturbed topography after the tsunami recovered to the stable one in the period from 2014 to 2015.

 Precipitation Radar (PR) onboard the TRMM satellite was operated for more than 17 years and the standard algorithm has been developed up to version 7 (V7). The standard algorithm version 06 originally developed for GPM’s Ku-band radar (KuPR) is ported to be applied to PR measurements and it is treated as PR standard algorithm version 8 (V8). In this study, rain rates R estimated by V7 and V8 are compared. (1) calculation of measured radar reflectivity factor Z_m, (2) attenuation correction and (3) conversion from attenuation-corrected radar reflectivity factor Z_e to R are considered as factors of the differences in R. (1) In V8, Z_m is higher because of the recalibration of PR, which increases R by about 17%. (2) attenuation correction does not affect the average value of R as siginificantly as (1) and (3). (3) In V8, as ε estimated by attenuation correction is generally smaller than V7, and smaller R is resulted from Z_e. On the other hand, R more than several ten mm/h is estimated more frequently in V8, because the upper limit of ε is not used.

 It is a significant problem on the disaster prevention to describe cloud developing process characterized by nonlinear property and identify the mechanisms of self-organization of cumulonimbus. In this research, we discuss the characteristics of vorticity that consider thermodynamic effects and its relationships between precipitation at the cross-section of convection core in the developing clouds using the isolated cloud simulation. Applying Fourie analysis to the vorticity field, the increasing tendency of the energy spectrum was shown. It also appeared that the behavior of the energy spectrum is different between our vorticity and the vorticity that is generally used. The comparison of the vorticity field and precipitation indicate the effect of latent heat caused by the generation of raindrops and graupel can affect the energy spectrum of the vorticity. We also applied the analysis to the developing clouds in the simulation of the past heavy rainfall event. The energy spectrum of vorticity shows the same tendency as the isolated cloud simulation, and our method has the potential to describe the characteristics of the triggers for cloud development.

 The first aim of this study is to investigate the features of diurnal variations in heavy rainfall frequency, and to relate them to typical meteorological fields during a warm season in Kyushu, Japan, using Self-Organizing Map (SOM). The result showed a tendency of higher heavy rainfall frequency in the morning compared with in the afternoon. The meteorological field patterns are characterized by the inflow of warm and humid air into frontal area. The second aim of this study is to clarify significant relationships between heavy rainfall hours of the day and disasters cases which occurred in Kyushu during 2006 and 2020 using the first and second maxima of heavy rainfall scale and hourly rainfall recorded in each hour. One of the trends is characterized by widespread and local heavy rainfall in the morning including sleeping time and in the daytime, respectively. Actually, the time-dependent features are true of disaster cases including 2017 and 2020 disaster cases in Kyushu. Therefore, the findings provide strong evidence for night-time disaster prevention in Kyushu.

 In this study, we analyzed the water vapor structure around convective clouds and tried to index the water vapor at middle-level through the results of numerical weather model simulations. First, the water vapor structure around the developed convective cloud was investigated, and it was shown that the water vapor content at an altitude of 2 km was remarkably large. The origin of this was that water vapor lifted by other convective clouds that preceded it continued to maintain a high humid for more than 40 minutes after the preceded clouds faded. Furthermore, the statistical analysis led to the interpretation that the water vapor content at an altitude of 2 km has a significant effect on the intensification, not the occurrence, of rainfall. Finally, a new index,AWV₁₊, was proposed to describe water vapor at an altitude of 2 km in anticipation of the operational status of water vapor observation technology in the next few years. Compared to the available precipitable water vapor index, AWV₁₊ was shown to discriminate the enhancement of rainfall intensity.

 Since March 2006, the Japan Meteorological Agency (JMA) has been operating Radar/Raingauge Analyzed Precipitation (R/A), which is updated every 30 minutes, as the basis of meteorological information for disaster prevention. Since July 2017, the JMA has been operating Immediate Radar/Raingauge Analyzed Precipitation (Immediate R/A), which is updated every 10 minutes. However no study has been conducted to analyze runoff using Immediate R/A. In this study, we quantitatively verified the accuracy of Immediate R/A using AMeDAS nationwide. At the same time, we evaluate the usability of R/A and Immediate R/A in flood forecasting by clarifying their effects on runoff simulation. As a result, the comparison with AMedas showed errors in rainfall nationwide. In addition, when Immediate R/A was used for runoff calculation, the runoff response will be more sensitive using Immediate R/A than R/A for rivers with smaller basin areas, but the effect on the runoff waveform was smaller for larger basin areas. In conclusion, Immediate R/Aof the rainfall products used for flood forecasting was found to be effective from the viewpoint of short distribution intervals, although its quantitative performance remained an issue nationwide.

 Extreme value (eg. annual maximum of daily precipitation) has bothered us by the surprising statistical behavior. A new record sometimes indicates an extraordinary value of return period, and a few top ranked observations are often found to deviate from the alignment of other order statistics. We have not been convinced by these results even if they are done by the orthodox methods of extreme value analysis. This study employ the two mutually connected statistics: simple extreme variable, which is equivalent to return period but treated as a statistical variable, and exceedance number for a fixed or statistically fluctuated threshold, and we explain explicitly the above-mentioned properties of extremes. These mathematical backgrounds of these two key concepts will enhance our data analysis and those interpretations of hydrological extreme statistics.

 Areal reduction factors (ARFs) and spatial extreme models are utilized in the analysis of spatial extreme rainfall. ARFs are ratios of the average amount of extreme rainfall in regions to the amount of extreme rainfall at the central point of their regions and are expressed as a function of their region sizes. The spatial extreme model represents the spatial correlation of extreme rainfall as a function of distances between points. The relationship between the two methods was confirmed previously based on the analysis of rain gauge data that were sparsely observed in space. This study reassessed the relationship between the two methods empirically utilizing radar rain gauge data in the Yoneshiro-gawa basin, Akita Prefecture that have high spatial resolution and confirmed that the relationship was not clear. The detailed spatial resolution data revealed that the estimates of extreme rainfall at points differed considerably although their positions were close to each other and the estimates of ARFs were unstable despite the similar spatial correlation structure estimated by spatial extreme model.

 Torrential rains which cause disasters seem to have increased and become severer recently. Extremes of observed precipitation at a location or catchment averaged one have been used for extreme analysis. There are two methods of extracting extremes from a time series for extreme analysis; annual maximum series(AMS) and peaks over threshold(POT). AMS has been widely used from its simplicity, however, POT used practically less than AMS because of its complicated process in threshold selection and POT extraction. This paper proposes a new threshold selecting method using L-moment and shows several notes in practical use. The result shows it estimates more stable return level with accumulating observation than GEV with AMS.

 Horizontal coordinate axis in a wave-current coexisting field like an inner bay is important in order to estimate Reynolds stress. Therefore, effects of coordinate axis on Reynolds stress depending on the directions of waves, mean flow and tidal current were investigated by a coordinate transformation of simultaneous horizontal velocities and eliminating wind wave components through a spectrum analysis. As the result of this study, it is made clear that an absolute value of Reynolds stress has to be small by a coordinate transformation to a wave direction in cases wind waves are dominant, and that to a mean or a tidal flow direction in cases wind waves are small. In addition, the approximate value of a dimensionless parameter to estimate the necessity to eliminate wind wave components from Reynolds stress was obtained. Therefore, detailed flow conditions would be understood by an appropriate horizontal coordinate axis in consideration with dominant directions of wind waves, mean flow and tidal current.

 We focus on an undular bore of tsunami run-up in rivers. To clarify an amplification characteristic of the undular bore in meandering rivers, we conducted the basic experiment and calculated it using the two- or three- calculation. The experimental results of wave height show that the amplification characteristic along the outer bank is different depend on channels. The longitudinal tendency of amplification is expressed by the two-dimensional calculation results. The three-dimensional calculation results corresponded to experimental results of the initial wave height. The experimental and calculation results showed us that the soliton fission amplification of the meandering channel of small sinuosity is larger than that of big sinuosity and the longitudinal increase of the wave height along the outer bank. We confirm the bore impinging on the outer wall of meandering channel and the wave height along the outer bank increased. For that reason, the soliton fission amplification occurs. Both calculation results are coincident with the profile of bore head’s line in meandering channels and can express the mechanism of the soliton fission amplification. The soliton fission amplification along the outer bank occurs in the meandering channel of small non-dimensional width. The bigger the sinuosity is, the larger the soliton fission amplification is.

 Aiming for the grasp of behavior characteristics of edge bore formed by tsunami, propagation and amplification/attenuation of an obliquely incident strong bore with two-dimensionality in plane space over a uniformly sloping beach are examined through the consideration and theoretical analysis of basic equations derived for the wave ray method for strong bores. As the results, it is indicated that 1) the bore propagation velocity depends on the static water depth h_0fc in front of bore, ratio h_fc/h_0fc of the total water depth h_fc of the bore front to h_0fc, bottom slope tanβ in the direction perpendicular to the bore propagation direction and slope dΔ(y)/dy of the bore height distribution Δ(y), 2) the straight propagation property of strong bore is higher than that of water wave except for some conditions, 3) the amplification/attenuation factor dh_fc/dt of h_fc depeds on h_0fc, h_fc/h_0fc, tanβ, bottom slope tanγ and water surface slope tanδ in the bore propagation direction and dΔ(y)/dy, 4) in the case that dΔ/dy=0 and bore is strong, analytical solutions of h_fc are derived and the larger tanγ and tanδ are the higher amplification height of h_fc becomes, and so on.

 The tsunami-induced bottom boundary layer is investigated based on the actual waveform obtained by the GPS buoys along the coast of the Tohoku region during the 2011 Great East Japan Earthquake tsunami. The k-ω model is utilized for the numerical analysis in this study. As a result, the boundary layer thickness was found extremely thin compared to the water depth. The velocity distribution was similar to that of the bottom boundary layer under wind-generated waves. The flow regime is located in the transition region from smooth turbulence to rough turbulence. Because of this, the bottom friction is underestimated by the steady-state friction coefficient, such as the Manning formula. Therefore, a simple method for calculating the bottom shear stress due to an irregular tsunami which has been proposed in the previous study is employed to make a comprehensive investigation.

 Satellite radar-based snowfall estimate is an essential source of information to understand cold season and high latitude hydrological cycle. In this study, snowfall detection performances of the satellite-based radar products (i.e., GPM-DPR, KuPR, and Combined Radar-Radiometer (CMB) retrieval algorithms) are evaluated against ground-based radar observation in Finland. It is found that the probability of detection (POD) is less than 50% when the snowfall rate is weaker than 0.7 ~0.8 mm/h. DPR and KuPR products show almost the same POD, and CMB product shows slightly better POD than other two radar products, presumably due to the added information from passive microwave observation. Overall, only 14 ~ 17% of the snowfall pixels are detected by the satellite radar products.