The accuracy of climate models decreases drastically when predicting rainfall for more than two weeks, so it is expected that deep learning models can be applied. In this study, deep learning is applied to the Chao Phraya River basin in Thailand to predict rainfall for one to three months during the rainy season. We examined whether the lack of observational data can be solved by using the CMIP5 dataset, which is a non-observational dataset. The results show that CMIP5 can be adapted to the training data. Furthermore, we found that deep learning methods tend to outperform climate models when the prediction lead time is long.

 In recent years, there has been a need to improve the accuracy of precipitation prediction using deep learning due to the increasing number of heavy rainfall events. However, due to the lack of spatial resolution, meteoro-logical factors other than precipitation were not learned. Therefore, we constructed a virtual observation point and investigated whether meteorological factors other than precipitation contribute to the learning process when the spatial resolution is increased. As a result, we found that meteorological elements other than precipitation are noisy and worsen the prediction accuracy in case of low rainfall, but improve the prediction accuracy in case of high rainfall. In addition, this tendency became stronger as the spatial resolution was increased.

 Prediction of discharge, water temperature and water quality in a river is necessary in evaluation of climate change impact in a coastal region. In this paper, we attempted to develop an evaluation model of hourly river water temperature by Artificial Neural Network (ANN) as an Artificial Intelligence (AI) technique. Since river water temperature is time-series data, the Recurrent Neural Network (RNN) with Long Short-Term Memory (LSTM) is adapted. As a result, a highly accurate prediction was shown using approximate one-year dataset as a learning data. Future prediction of river water temperature under the present and the future projected climate conditions was conducted in the Yabe River by combining with d4PDF and hydrological model. From the result, utilization of the developed model was clarified.

 Prediction of snowmelt water in early spring is very important for water use and disaster prevention managements in heavy snowfall regions. Generally, current snowmelt runoff models have physical processes, and the parameters are calbralted by trial and error due to some uncertainties such as snowfall estimation. Therefore, we estimated the inflow of resourvoir during the snowmelt period by using deep learning for the Sagurigawa dam and the Oishi dam in Niigata prefecuture, which are one of the most heavy snowfall areas in Japan. The inflow were calculated by combining three deep learning model and four input datasets in this study. The reporoducivillties of inflow by deep leraning models were well in normal and heavy snowfall years. However its accuracy decreased in light snow years. Futhermore, it was suggested that the accuracy of the Recurrent Neural Network - based learning model with input data period of 10 days tends to be high, and the air temperature as the input data greatly affected the accuracy.

 Recent rash of severe flooding events have increased the number of victims due to delayed escapes in Japan. To reduce the victims, it is necessary to develop a system to predict the water levels and discharge efficiently at multiple points in a river basin. In this study, for the Saba River in the Yamaguchi Prefecture, we built a deep-learning model with LSTM to accurately predict water level and discharge 3 hours ahead. From the analyzed results, it was shown that the water level 3 hours ahead can be predicted with high accuracy at multiple points in the basin. On the other hand, the logarithmic conversion was applied to the discharge values to improve the learning efficiency against data of high dynamic range. With this modification, the accuracy of the model prediction was greatly improved on the peak discharge rate and delay time of occurrence compared to the case without the logarithmic conversion.

 It has been reported that machine learning predicted high-water levels of flood events. In this study, several approaches of water level prediction were carried out based on machine learning models for the Chikuma River suffering huge damage in the typhoon No. 19 of 2019. Deep learning and linear regression models indicated the high accuracy of the prediction of extensive high-water levels under the limited condition. Furthermore, it was suggested that AI predicted the incremental water levels based on the water level data of observation stations, and is able to predict flood waves at high accuracy, complementary applying the flood peaks and the starting time of water level rising acquired from accumulated rainfall into the prediction.

 Streamflow data are important for river maintenance, water resources planning, and flood forecasting. However, the flow data may be missing due to various reasons. Therefore, this study proposed a novel method using deep learning to complement missing data of flow discharge time series. This study utilized hourly flow data before and after missing and precipitation data as input. As the deep learning methods, this study selected 1D CNN and MLP. The results showed the high capability of 1D CNN and MLP. Especially, 1D CNN was able to estimate the missing flow rate well. In addition, it the results in this study indicate that it is important to use precipitation data in addition to flow discharge data as input in order to accurately estimate missing flow discharge.

 Typhoon Hagibis (2019) caused severe rainfall damage mainly in the Kanto region. It is necessary to evaluate the maximum possible rainfall for the Kinki and Chubu areas which have been directly unaffected by Hagibis, with taking into account both the effects of worst track and global warming. In this study, we perform pseudo-global warming track ensemble experiments using a high-resolution typhoon model to evaluate the maximum possible rainfall under the worst-track scenario and under the future global warming scenario (during 2080-2099 in RCP8.5). The future changes of the maximum possible rainfall are +200 mm in the Kii Peninsula and +50－150 mm in the upstream areas of the three major rivers in the Chubu area. The future changes in the maximum possible rainfall due to global warming is comparable to the difference of accumulated rainfall caused by an east-west deviation of about ±0.5° from the worst-case track.

 A high-resolution pseudo global warming experiments were conducted for line-shaped convective system, 2012 Kameoka event and 2014 Hiroshima event, under the RCP8.5 and RCP2.6 scenarios. For 2012 Kameoka event, line-shaped rain band was gradually intensified while for 2014 Hiroshima event lineshaped rain band gradually disappeared as the warming scenario became higher. This was due to the fact that the weak rainfall area upstream of 2014 Hiroshima event was enhanced antecedently by the warming effect, and disrupted the environmental field for the occurrence of the line-shaped convection system as expressed such as CAPE and bulk Richardson number. It is indicated that line-shaped convective system is generated when thermodynamical and dynamical various factors are well balanced, and the effects of ambient disturbances should be taken into account when we apply pseudo global warming experiment.

 The impact of climate change especially has been becoming apparent after 2018 Japan floods, was reported as included the effect of global warming for the first time. Therefore, it is important to consider the impact of climate change in future flood control plan. However, in the previous studies, the timing when the impact of climate change begun to appear has not been revealed based on observational meteorological data. Accordingly, this study aims to clarify the timing when the impact of climate change begun to appear based on CCI (Climate Change Indices) with precipitation of meteorological office data in Japan. As the result, we confirmed that the timing when the impact of climate change begun to appear may be around 2014 in whole Japan. Additionaly, it is found that the timing is vary spatially.

 In this study, the relationship between extreme rainfall and near-surface temperature was investigated by using ensemble climate data based on the scaling factor (SF), which is the rate of increase in rainfall intensity per 1 °C increase in temperature. Although the observed SF generally follows the ClausiusClapeyron relationship (7 %/°C), it is found to vary from 5 to 10 %/°C among sites. The factor of this variability was investigated by using ensemble climate data, and it was found that this variability is caused by the number of observation samples. The impacts of climate model resolution and climate change on SF were also investigated. These results indicate that a high resolution and a massive number of samples are required to discuss the SF of extreme rainfall. The results also suggest that short time heavy rainfall risk at high temperatures will increase because the SFs are higher under the warmer climate.

 With the growing importance of taking stepwise adaptation measures to climate change, we conducted a future change analysis of rainfall during the Baiu season using 150-year continuous run. We analyzed the future changes of baiu frontal zones using average daily rainfall and baiu extreme rainfall using hourly rainfall. As a result, the rainfall area gradually moves northward in both the frontal zone and location of the extreme rainfall events. However, there were also regions that contained periods when rainfall was lower than the historical average while there was a marked tendency for rainfall to increase in northern Japan. This reasons are discussed in terms of the influence of sea surface temperatures and typhoons. Furthermore, we showed that the duration of extreme rainfall events will gradually increase in the future.

 This study discusses the case in which the maximum annual daily precipitation is larger in the 2-degree warming experiment than in the 4-degree warming experiment in the future projection by large ensemble dataset. The results show that such inversions occur in about 20% to 30% of the Japanese area in some experiments. One of the reasons is that the difference in results between two warming experiments is smaller in the SST pattern where many inversions occur than in the other SST pattern experiments. In order to avoid such an inversion problem, a two-pass bias correction method is presented. However, it is important to note that the bias correction can result in a correction of up to 50% before and after the correction.

 In this study, we proposed a method to make a composite probability by using the total rainfall used in the hydraulic control plan and the cumulative rainfall in time that contributes to the peak discharge of the river. The rainfall probability was calculated using the rainfall data of the past experimental results of the meteorological enseble data “Database for Policy Decision-Making for Future Cliate Change”. The rainfall time that affects the peak discharge was set from the correlation coefficient between the peak discharge of the runoff calculation results using the past experimental results and the cumulative rainfall at each time(1hour to 72hour). We evaluated the rainfall with a compound probability and found the rainfall that caused flood damage with a higher probability of ozccurrence than the actual rainfall.

 The recently developed dual-polarization phased array weather radar (MP-PAWR) is capable of spatio-temporal high-density 3D observations and early detection of localized heavy rainfall compared to current weather radars. The purpose of this study is to develop a flood forecasting method that can contribute to evacuation and lead time during localized heavy rains by using the highly accurate observation information of MP-PAWR. In order to achieve this goal, we calculated the predicted rainfall based on the temporal variation of the vertically integrated rainfall, and converted the predicted rainfall into the river water level using a rainfall-runoff model. As a result, it is shown that the forecasting model can be applied to estimate the rise time, peak time, and peak water level 30 minutes before the actual peak water level occurs in the target rainfall event.

 Thermal Image Velocimetry (TIV) is the method to measure the 2-dimentional wind field tracking the temperature displacement caused by advected turbulent heat exchange using TIR camera. So far, this method has been applied to the surface whose heat capacity is low so as to easily capture the turbulence-caused temperature displacement. We applied TIV to the high-heat-capacity asphalt road within urban canopy for the first time. As a result, we could conduct TIV observation even to the asphalt surface, observing complex wind field within the uraban canopy. Then we need to conduct temporal filtering with larger filter than the cycle of the random noise, which can be estimated by the spectral analysis of the displacement of the surface temperature. Even low performance TIR camera can conduct TIV observation if we set the large temporal filter enough compared to that of high performance TIR camera. And we found the possibility that we can change the size of tracking turbulent structure by changing the filter size,

 The objective of this study is to observe and analyze the Doppler lidar signature before and during the first echo from Ka-band radar. In addition, two time-lapse cameras are also utilized to capture the visual evolution of the cloud. Radiosonde from the nearest launch site is also used in this study for meteorological data. The observation data used in this study is a case study of an isolated convective cumulus cloud on August 18, 2019. Ka-band radar captures the first echo at 14:34 JST during this period, confirmed by the cloud image from a two time-lapse camera. Updraft exists at the bottom of the first echo. This presence is strengthened by convergence value, doppler velocity variance during the first echo and quantitative result between this divergence-convergence and variance in 30 minutes before the first echo. The Doppler lidar also detects pairs of vorticity before and during the first echo. This pair of vorticity is lifted vertically, known as the vortex tube, and is moved towards the northwest due to the environment wind in the lower levels. A potential reason why vortex tubes exist is the condition of vertical shear added with the updraft presence. This small vortex tube is located at the bottom of the baby-rain cell cumulus cloud and could only be detected using the sensitive Doppler lidar.

 Simulation of seven sea-breeze days (SBD) during dry season in tropical megacity of Jakarta was carried out using WRF with detailed urban representation. Model simulations were evaluated using satellite derived cloud line and high-temporal-resolution meteorological data obtained from observation campaign in 2017 and 2018. Results shows that WRF with detailed urban representation was able to simulate sea-breeze features and the associative boundary layer development. In the early stage of sea-breeze, model convergence line associated with sea-breeze front were matched against cloud line derived from satellite imagery. WRF tend to produce earlier sea-breeze occurrence due to overestimation of shortwave radiation and underestimation of latent heat flux. In general, simulated wind speed, temperature and relative humidity shows good agreement with observed values. Model also able to well simulate sea-breeze features, including lower boundary layer over Jakarta associated with thermal induced boundary layer (TIBL). Sea-breeze TIBL is influenced by coastal form of Jakarta and might plays important factor in air quality over the city.

 This study investigates an analytical solution of downslope windstorms to the non-shear and continuously stratified flow. Based on a previous study (Winters and Armi, 2012), which describes a upstream blocking and changes from a subcritical to a supercritical flow, we formulate the internal jump over the lee slope. Furethemore, we extend this theory to the no upstream blocking condition and examine the applicable range of this theory in comparison to previous regime diagram (Lin and Wang, 1995).

 Japan has suffered from devastating flood disasters caused by localized heavy rainfall known as Guerrilla heavy rainfall (GHR) recently. For reducing the vital damage to human life and property, it is necessary to predict the risk of heavy rainfall precisely. To better alert the risk triggered by GHR, we aim to propose an advanced quantitative risk prediction method in this research. One of the improvements was that the relationship between the predicted risk level and the variables (i.e. the reflectivity and vorticity) was considered depending on each rain stage because the variables showed different characteristics according to the development of the convective cloud. The other one was that the variables (i.e. the vorticity, divergence, and updraft) were estimated with real wind field data by multiple Doppler radar analysis. Then, the multilinear regression was used for finding the correlation between the predicted risk level and the variables with accuracy. The accuracy of multilinear regression was estimated by a Receiver Operating Characteristic analysis. As the result, the most appropriate regression among the relevant variables was composed of reflectivity, vorticity, divergence, and updraft by multiple Doppler radar analysis. It is possible to predict the risk quantitatively with high accuracy of 90% at the early rain stage.

 This study aims to evaluate the reasonability of future probable maximum precipitation (PMP) values estimated by the moisture-maximization method. This study utilized the d4PDF database, which is a large ensemble climate simulation output, to evaluate the estimated PMPs with sufficient extreme precipitation cases. Using the d4PDF, PMPs were estimated under historical and future climatic conditions in four target areas of Japan. The estimated PMPs in each climate condition were evaluated for their reliability with reference precipitation estimated from the annual maximum daily rainfall of d4PDF. The rate of change of PMPs between the two climate conditions was lower than that of precipitation in all areas. Compared to the reference values, historical PMPs were reasonably estimated while the future PMPs were underestimated. The future moisturemaximizing ratio (MMR) was estimated to be lower than the historical MMR as the event precipitable water (PW) had a higher change rate than the maximum PW in future climatic conditions. The lower future MMR led to a low change rate and underestimation of PMP in future climatic conditions compared to the historical ones. Consequently, the moisture-maximization method may have an uncertainty in underestimating the PMP under future climatic conditions. Therefore, it may be necessary to modify the moisture-maximization method for suitable PMP estimation under future climate conditions.

 We developed an index of heatstroke risk by using the database of heatstroke patients provided by Saitama Prefecture. The index is based on the relationship between the physiological responses of the patients (body core temperature, skin temperature, amount of sweating and blood flow) and cumulative percentage of the patients transported at each level of the physiological responses. The physiological responses were estimated by using meteorological data, an urban canopy model and the heat balance model of human body, considering the individual vulnerability for heatstroke. The indicator was tested with an actual heatstroke patient transported at night in the summer of 2018. Even though the environment where she stayed was not typical of heatstroke, the indicator showed the alert level for heatstroke.

 In order to verify the realistic temperature reduction effect of cool roofs, we conducted a more realistic weather simulation than previous studies, using a WRF model that reflects the urban geometric parameters of a real city, and varying the reflectance and diffusion rate step by step. In the case of cool roofs installed only in the 23 wards of Tokyo, it was confirmed that the amount of temperature reduction during the summer season increased as the reflectance and diffusion rate increased, and that the effect extended beyond the installation area. When the reflectance of all roof surfaces was 85%, the daily mean temperature decreased by about 0.32°C and the daily maximum temperature decreased by about 0.38°C. The correlation between the area of the cool roof and the temperature reduction was confirmed, indicating the importance of setting detailed urban geometric parameters.

 Operational weather forecasting centers in the world are vigorously trying to utilize cloud and rain signals within satellite microwave observation for improving weather prediction. However, there is still no established technique that is effective over land where the background radiation is strong. This study developed a coupled atmosphere and land data assimilation method to fully utilize Ka-band microwave (KaMW) including relatively weak cloud signal for local-scale rainfall prediction. Results of its application to a heavy rain event showed that assimilation of KaMW predicted more appropriate rainfall amount than that of high-frequency microwave (HiMW), which contains large ice signal, while the method with HiMW more effectively predicted rainfall area. This indicates a large impact of KaMW on rainfall prediction and enables reducing the information requirements for cloud estimation over land and using HiMW for more detailed estimation and prediction of cloud and rain in the future.

 Land surface modeling is important for hydraulic engineering and water resources management. To improve the accuracy of land surface models, it is necessary to optimize their unknown parameters and assess their uncertainties. However, the multi-objective parameter optimization and uncertainty assessment, which realizes to improve all relevant state variables such as soil moisture, vegetation dynamics, and surface temperature, are yet to be developed. We developed the method of multi-objective parameter optimization with uncertainty assessment by multiple satellites’ observations. Soil moisture, LAI, and land surface temperature are chosen as land surface conditions, and optimization was done in order to simultaneously decrease the 3 types of errors between the observation and the corresponding model’s variable. As a result, the RMSE between the observations and the model’s variables are generally decreased. Also, it was revealed that ecologilcal observation has the strongest impact on the improvement of the model’s accuracy.

 Frost heaving of soil layers damages structures in and on earth surfaces. Therefore, deformation processes due to freezing and thawing of the soil layers have been widely investigated. X-ray computer tomography was applied to a silica sand layer to nondestructively and three-dimensionally visualize the deformation process. The resultant images were analyzed by a digital volume correlation method to calculate the displacement and the volumetric strain of the sand layer. The deformations due to expansion and shrinkage of water and ice portions in the upper soil layer were detected as the horizontal and vertical plane distributions of the volumetric strain before and after freezing and thawing.

 Ice jam causes a sudden rise in water level and an accident involving the flowing river ice, which is one of the problems in the maintenance of the frozen river.We aim to develop methods for predicting the location and occurance time of ice jam to reduce the damage caused by ice jam. From the viewpoint of practical use, it is necessary to examine series of prediction methods under appropriate procedure and conditions. We organized the examination procedure considering a series of processes from the start of deicing of river ice to its occurrence, and the flow, destruction, and deposition, and examined its applicability. We examined prediction process and results from the viewpoints of (1) appropriate pre-extraction of areas where ice jam will occur in each river channel, (2) appropriate pre-prediction of the start time of deicing, and (3) appropriate pre-prediction of the time when the risk of ice jam occurrence increases.

 Accurate prediction of dam inflow during snowmelt is important. Snowfall on mountain slopes is redistributed by wind, gravity, and topography. However, this effect is not taken into account in general snowmelt runoff analysis. In this study, based on the results of previous studies, snowmelt runoff analysis considering snow redistribution was conducted for the Chubetsu Dam basin. As a result, it was confirmed that the constructed runoff analysis model had good reproducibility, and that the reproducibility could be improved by performing snowmelt runoff analysis that takes snow redistribution into account. In particular, the reproducibility of peak values in the latter half of the snowmelt season was improved.

 This study evaluated the changes in the exposed population to river flooding, inland flooding, and slope failure from 2015 to 2100 with population changes by shared socioeconomic pathways (SSPs). The exposed population to river and inland flooding was defined as the population living in the area with inundated depth of 45 cm or more, which expressed over a floor level of general houses. The exposed population to slope failure was defined as the population living in the area where the probability of slope failure exceeds 80 %. Although there were differences in population distribution among the SSPs, such as urban concentration and decentralization, the magnitude relation of the total exposed population to the three disasters by SSPs was the same as the magnitude relation of the total population by SSPs in 2100. The SSPs with lower population concentrations in high-hazard areas than other SSPs were estimated to be SSP3 to river flooding, SSP4 to inland flooding, SSP1 and SSP2 to slope failure, and SSP2 and SSP3 to the three disasters.

 We analyzed the human damage situation caused by the recent flood disasters in Japan and developed a new loss of life estimation method for flood risk assessment. The relationship between fatality rate and inundation situation, which are represented by inundation depth and water rise speed, was analyzed. In the analyzed disasters, fatalities occur only when water depth is higher than 1.5-2m and fatality rate is high when water rise speed is rapid. The developed model estimates loss of lives for each class defined by water rise speed, timing of inundation and age, of which fatality rate are obtained based on the observed fatality rate. The applicability of the model was examined by applying to the flood disasters in Japan and the result shows that the model can accurately estimate loss of lives caused by the flood disasters in Japan.

 “Robustness, Resilience after damage and Emergency response” are keys to achieve disaster resilient society. Torrential rainfall in July 2018 affected the enterprises in wide area of the west Japan, bringing direct loss due to inundation and landslides, and indirect loss. This paper conducted an questionnaire survey to the affected enterprises in Hiroshima and Okayama Prefectures to understand the actual situation of direct/indirect loss of their headquarters, branch offices and factories. It was found that many enterprises suffered business disruption due to indirect loss such as water outage, road brockage and logistics disruption.

 This paper presents the simulation of the flood by Typhoon 19 October 2019 using two flood models. Firstly, the discharges around Tokigawa and Oppegawa rivers were reproduced by a distributed rainfall-runoff/flood-inundation model. The simulated discharges were in good agreement with observed discharge. Then, the inundation simulation with 5 m resolution elevataion data was carried out using Fugaku super computer. The detailed inundation process including the effect of second dike was reproduced well which shows the effectiveness of fine resolution simulation. Finally, the effect of the climate change on the catchment average 24 hour rainfall over Arakawa river was investigated using d4PDF. The 24 hour rainfall–return level was plotted using the data which shows the increase of the rainfall intensity toward the future.

 As a result of the studies that have been carried out in the Hokkaido region triggered by the heavy rains in Hokkaido in August 2016, a method for studying hydraulic control plans that takes climate change into consideration has been proposed. However, long-term investment is required to achieve the planned hydraulic control plan with a limited budget, and efficient investment is required during the investment period. In this study, by applying the modern portfolio theory to the uncertainty of damage reduction effect of river improvement caused by rainfall temporal spatial distribution, it became possible to discuss the efficient investment ratio for river improvement of each river in the basin using rainfall temporal spatial distribution obtained from a large amount of ensemble data.

 The purpose of this study is to examine the potentially high risk zones where the horizontal evacuation of the living people would be required before and during the flood inundation situations. In order for that, the effects of the buildings in flood inundation simulations on the results with the derived horizontal evacuation zones were evaluated based on the flood inundation simulations in this study. The flood inundation simulation was applied to the Chikuma River, which was flooded by the heavy rains of Typhoon No. 19 in 2019, and several methods taking account for the buildings in the flood inundation simulation are examined. In addition, flood inundation simulation considering buildings and risk classifications are applied to the Jinzu River in Toyama Prefecture, and vertical and horizontal evacuation zones are evaluated. As a result, the horizontal evacuation zones in Toyama city was successfuly evaluated. In addition, we proposed a hazard visualization method in 3 categories including inundation damage prevention areas.

 The purpose of this study is to verify the function of a removable stream fishway installed in the head construction of an urban river. Therefore, we carried out the experiments in the field where is Koedo River in Wako, Saitama Prefecture. In the experiments, it was used juvenile Ayu; Plecoglossus altivelis altivelis of body length 4.0 cm to 5.8 cm which number of fish is 100. As the results, approximately 8 % of juvenile Ayu run up to most upstream of fishway and approximately 24 % of juvenile Ayu to went upstream from the entrance of fishway, despite the short experiment time of 15 minutes. Therefore, it was confirmed that the removable stream fishway installed in this study. If the time is long, it is thought that juvenile Ayu can run up more. In addition, many swimming fish can be expected to use this fishway. Moreover, the fish collection situation was fair, because juvenile Ayu at 50 % of the probability were swimming toward the downstream entrance of the fishway. It is advisable to provide priming water and others to collect more fish, however it is considered to be strict in terms of conditions. Therefore, the expectation of fish collection can be improved by providing measures to prevent invasion suitable for this field.

 One of the reasons for the decline of eel population is the discontinuity of water level caused by dams and weirs, and the installation of fishways is one of the remedies. Previous studies have attempted to elucidate the migrating characteristics of eels by varying the arrangement, diameter, and spacing of the cylindrical protrusions installed on the bottom of fishways for eels. However, few studies have been conducted by varying the total length. In this study, the spacing of the cylindrical protrusions was varied from 7.5 to 50 mm, and the total length of Anguilla japonica was varied from 150 to 300 mm to obtain the optimal spacing of the protrusions for the upstream migration. As a result, it was clarified that the migration rate of eels with a total length of 150-300 mm peaked at around 15-30 mm as the spacing increased, and then decreased at 50 mm. In the space of 15-30 mm, the stagnation time and the frequency of temporary return to the downstream decreased with an increase of the space, and it was found that the eel migrates linearly. Furthermore, an empirical formula for evaluating the ground speed of migration based on the spacing between the protrusions and the total length of the eel was suggested.

 Fish has body temperature synchronized with the surrounding water temperature. Several studies have reported that some fish are killed by cold water caused by the rapid shutdown of industrial wastewater. It was known that rapid water temperature changes have a large effect on swimming behavior of fish. However, the swimming behavior under the condition of rapid water temperature changes has not been investigated so many. In this study, an investigation on swimming behavior of Opsariichthys platypus under the condition of rapid water temperature changes was conducted. It was found that the tailbeat frequency, swimming distance and swimming speed increases with an increase of surrounding water temperature as compared with the acclimation temperature. Further, it was suggested that fish feels fear and increases swimming speed when fish meet 10℃ high temperature water than the acclimation temperature.

 In the Asahi River in Okayama Prefecture, the Heidan area located at the tidal zone is home to natural spawning habitats for ayu. Recently, effective development of spawning grounds has been practically required for the resource multiplication in the Asahi River. However, there are some concerns on un-known factors of salinity intrusion and related physical phenomena. In this study, we conducted field observations and numerical simulations for clarifying the suitable spawning environment for ayu in the Heidan area. Results showed that the optimal area for spawning in the tidal zone is characterized by the absence of salinity water throughout the targeted investigation period. Results also demonstrated that the riverbed granularity and flow conditions arising during spawning hours at night range approximately within the preferred values examined in earlier works, which have been conducted under the unidirectional river flow condition without a tidal effect.

 Smallmouth bass have been increasing in Japan in recent years as invasive alien species and exterminated in invaded river and lake. This study evaluated the year until eradication and the equilibrium population of smallmouth bass using population model developed in the Kankakee River. When controlling eggs, juveniles, and adults independently, A high removal rate(74%－98%) was required to achieve the eradication rate of 100%. when adult and juvenile were additionally removed under conditions where 80% of eggs were removed, the eradication rate could be increased to 100% with more than 20% less removal rate than when each growth stage was exterminated individually. Therefore, it is important to remove eggs, juveniles, and adults in parallel in actual extermination. By updating this model in the future, it is expected that the model will be applied to rivers in Japan and that the accuracy of estimating the number of populations to be removed and the number of years required for eradication and low-density control will be improved.

 This study aims to clarify the flood inundation conditions and vehicle loss at the Hatagawa Riner basin, where vehicle death occurred by typhoon Hagibis in 2019. In this study, we conducted 1D flow and 2D flood simulations in the Hatagawa River basin. The result indicated that a rapid increase in flow velocity and water depth were found at the disaster point of death in the vehicle. The simulated time when the vehicle was outflowed was almost comparable to the actual time. It is found that the area for vehicle outflow were locally limited and were in agreement with the inundation depth distribution affected by the topography of alluvial fans and railway embankments.

 In recent years, global climate change due to global warming has become remarkable, and one of the places where the most serious damage is considered to occur in the situation where record heavy rains occur frequently is the underground space.In particular, the subway that spreads out in central Tokyo has many users, and there is an urgent need to take measures against inundation.In this paper, we prepared the case where the external force is only rainfall and the case where the maximum possible case is prepared, and analyzed inundation risk in the target area centering on the central part of Tokyo.In addition, we investigated the amount of inflow water flowing from the road on the ground through the connecting point of the subway station and how it behaves in the subway tunnel, and clarified the inundation expansion process for each line according to each external force condition.

 For numerical analysis of canal network flow in the field of irrigation and drainage, the authors developed a method that supports one-dimensional (1-D) open-channel flow analysis, using a geographic information system or GIS. A developed computer program including this method enables users to: (1) draw multiple channels with a mouse on a GIS map, where connections between channels and agricultural fields are automatically recognized and the 1-D mesh-file is generated for the simulation model; (2) visualize the simulation results on a GIS map using shapefiles generated in (1); and (3) summarize predetermined water demands in agricultural fields based on the canal network structures. The method was evaluated by being applied to two study areas: an irrigation canal system, and a drainage channel system. To implement simulations, less workload is required using this method than without it, especially for recognizing the connection between each agricultural lot and canal node. With this method, all channels in a system can be treated as 1-D flow, taking each agricultural field into account.

 Oil spill accidents during the heavy rainfall and flood are reported. For the appropriate manipulation to the oil spilled to water area, it is needed to predict the temporal and spatial variation of oil spreading area.

 In this study, first, numerical simulation model of oil considering to the weathering processes of oil was developed, because the residential time of oil at sea area is becoming longer. Next, using this model, numerical simulation was carried out supposing the oil spill into enclosed coastal sea. Computed results shows that oil is decreased by the weathering and also changed to particulate component according the wave generations. Spatial distribution of oil slick at water surface is strongly affected by wind and its range is differ from other components that flow in water.

 In summer 2020, southern China was subjected to continuous heavy rainfall and the Yangtze river basin suffered severe flood damage. Understanding the causes and characteristics of major floods (including the 2020 floods) will greatly aid future flood management strategies. We analyzed the 2020 Yangtze river basin floods; satellite observations provided spatial and temporal rainfall, surface wetness, and flooding data. CHIRPS data from rain gauges and satellite observations were analyzed in terms of spatial distribution of intensity and amount of rainfall. The normalized difference frequency index (NDFI), derived via passive microwave remote-sensing, revealed the extent of surface wetness and flooding. The results of the analysis revealed that 1) high intensity of rainfall was observed mainly in Min and Jialing river basin in upstream, and higher anomalies of total rainfall was observed in both upstream and downstream 2) higher values of NDFI anomalies were widely distributed not only on the downstream but also on upstream 3) relatively larger anomalies of NDFI was found more on upstream in 2020.

 A superposed inundation simulation is conducted on the conditions, the sewer does not function and the probable rainfall is 1000 years, for a compound water disaster in which inland and foreland waters inundation occured at the same time, targeting the second-class river Gobo River in Takamatsu City, Kagawa, JAPAN. The results show that the inundation area increases and the flow velocity partly increases in the compound water disaster compared to the cases where the inland and the foreland inundations are occurred individually. Therefore, it is assumed that the difficulty of evacuation will increase, and it will be necessary in the future to consider superimposed inundation on the hazard map of water disaster.

 When the many sections of riverine levee systems are damaged due to a large-scale earthquake, it is necessary to allocate limited restoration materials and personnel appropriately, and ensure the safety of the area against floods as soon as possible. It is necessary to promptly estimate potential flood damage under the conditions where the levee system is damaged for deciding the priority of emergency restoration work sections. In this study, we developed a calculation program that can quickly estimate potential flood damage during an emergency restoration period while considering risk transfer from upstream to downstream. By test applying the program to a model area, it was shown that the restoration scenario which prioritizes the reaches where the flood risk reduction effect is large has a greater flood risk reduction effect than the scenario which prioritizes the reaches where the flood discharge capacity is small.

 In order to analyze the effect of flood experience on the decision-making criteria for residents' evacuation behavior, an agent model that simulates the acquisition of evacuation criteria by reinforcement learning from the flood experience simulated on a computer is proposed. The model repeats the experience where it starts to evacuate with reference to nearby river water level as an evacuation inducing switch and receives reward according to the appropriateness of its behavior. It obtains the evacuation criteria (threshold value of river water stage) which maximize the reward value. Through the learning process based on various rewarding system and various scales of virtual flood experiences, the most promising rewarding rule is to give positive value only when the agent reaches the shelter safely in flood cases where its house is flooded.

 This paper presents the results of an assessment of the My-TimeLine seminar to a voluntary disaster prevention organization and a flood defense party on their knowledge and thoughts and their evacuation behavior intentions about flood by a questionnaire survey. The results of this study are summarized as follows. 1) Making My-TimeLine significantly improved most of knowledge and thoughts about heavy rain and flood, and evacuation behavioral intentions after the seminar compared to before the seminar. 2) On the other hand, making My-TimeLine significantly didn’t reduce the psychological cost of coping with evacuation (evacuating to a shelter). 3) We confirmed that making My-TimeLine sufficiently improved risk awareness related to evacuation intentions.

 Several levees of the Oda River and three tributaries broke down in Mabi-cho, Kurashiki City due to the heavy rain in western Japan in July 2018, and the number of casualties due to inundation reached 54. The subsequent heavy rain caused by Typhoon No. 19 in 2019 and the floods caused by the heavy rain in July 2020 also flooded many rivers and caused many casualties. In order to reduce casualties in the event of such a large-scale flood, selecting an appropriate evacuation route based on the temporal changes in inundation depth and flow velocity due to flooding from inland waters and flooding from rivers, and early evacuation actions is important. In this study, the optimum evacuation routes were examined using numerical results of inland water inundation due to rainfall and inundation due to bank collapse in Mabicho, Kurashiki City, and evacuation sites and start time for safe evacuation were also examined. As a result, the necessity of adding an evacuation site and the safe evacuation start time were clarified.

 Flooding caused by heavy rains has occurred in many parts of Japan and therefore it is important to encourage voluntary evacuation behavior for damage control. In this study, we examined an evacuation simulation that applies the results of river inundation analysis using iRIC, and examined the effectiveness of vertical evacuation and voluntary shelters. The simulation results show that vertical evacuation is effective, and that evacuation to voluntary shelters increases the evacuation completion rate and reduces the time to complete evacuation. By examining multiple cases of breaches, the possibility of understanding the risk of evacuation in the target river is demonstrated.