Long-term climate change may strongly affect aquatic environments in mid-latitude water resources as well as flood and landslide disasters. Water temperatures in reservoirs have an important factor that controls the aquatic environments. In particular, surface water temperature strongly responds to air temperature that includes the impact of climatic changes. This study focused on temporal variations in surface water temperature in a mid-latitude reservoir (Tokachi Dam, Japan) that could be affected by long-term changes of air temperature in the future. In addition to conventional methods such as a linear regression, deep neural networks (DNNs) were adopted to understand the long-term relations between air temperature and surface water temperature. This is because the DNNs easily deal with nonlinearity data, including uncertainties, that are obtained in complicated climate and aquatic systems. However, the DNNs cannot predict appropriately for unexperienced data (i.e., no training data) such as future surface water temperature, which nobody experiences. To solve this problem, a transfer learning (TL) approach was introduced to the DNNs to predict reasonably even if the reservoir unexperienced. Observed data that obtained in the reservoir for 30 years were used to train a DNN-based model (DNN model) for the validation of the model. Continuous projection data (i.e., air temperature) ranging over 150 years for climate change, which were obtained from climate models, including a downscaling reginal model, were used as the input data of the DNN model to predict the past and future surface water temperature in the reservoir. The results showed that the DNN model with the TL approach predicted appropriately based on the difference between past and future air temperatures. The model suggested that the occurrences in highest water temperature increased and the occurrences in lowest water temperature decreased in the future.

In Japan, as Typhoon Hagibis of 2019 caused a lot of damage to houses due to inland flood in the eastern area of Japan, countermeasures against inland flood have become more important. There are concerns that inland flood damage will increase due to climate change. It is necessary to quantitatively evaluate the effect of adaptation measures. This study focuses on improving the capacity of inland water-drainage facilities as an adaptation measure and evaluates the damage-mitigation effect of the adaptation measure considering climate change in Japan. The inundation depths and the damage costs were calculated using a two-dimensional unsteady flow model and the Manual for Economic Evaluation of Flood Control Investment published by the Ministry of Land, Infrastructure, Transport, and Tourism. We assumed that rainwater in the inland area cannot drain into rivers due to the high river-water level. We also assumed that the higher capacity of inland water-drainage facilities would facilitate coping with precipitation with a 10-year return period from a 5-year return period. As a result, between the current (1981-2000) and near-future (2031-2050) climate scenarios, the expected annual damage cost (EADC) except for the adaptation measure increases by approximately 2.1 times in the representative concentration pathway (RCP) 2.6 and RCP8.5 scenarios. On the other hand, between the current and late twenty-first century (2081-2100) climate scenarios, the EADC except for the adaptation measure increases by approximately 1.8 and 2.5 times in the RCP2.6 and RCP8.5 scenarios, respectively. Further, by the higher capacity of inland water-drainage facilities, the EADC in future scenarios can be reduced by 28–37% compared to the EADC except for the adaptation measure. However, the EADC in future scenarios exceeds its current-climate when the capacity of inland water-drainage facilities is improved. This result indicates that only a drainage system cannot prevent future inland floods.

The risks associated with land use and climate change are a common concern, especially in relation to their potential impact on many cities around the world. Jakarta, Indonesia is a typical urbanized Asian city where flooding presents a challenge. Several factors contribute to floods in Jakarta, such as land-use change (urbanization) upstream, land subsidence, and climate change. This study quantified the impacts of land-use and climate change using a flood inundation model to analyze several future urban-growth and climate-change scenarios. A physical rainfall-runoff and flood inundation model was developed to understand the flood inundation mechanism in Jakarta and to investigate the impact of land-use and climate change on flood inundation. The model consists of a rainfall-runoff module for each sub-basin, a hydrodynamic module in the river and canal networks, and a flood inundation module for the floodplains. We investigated the effects of changes in land use on flooding in Jakarta by predicting future land use using the SLEUTH model, which estimates urban growth based on historical slopes, land use, exclusion, urban growth, transportation, and hill-shade data. For the urban-growth scenarios, this study used the RCP8.5-SSP3 (worst-case) and RCP2.6-SSP1 (compact-growth) scenarios. For land subsidence, we assumed that historical land-subsidence speeds in the target area varied linearly over time. The RCP8.5-SSP3 and RCP2.6-SSP1 scenarios were also used for the climate-change scenarios. A statistical method and regional climate model (WRF) were used to downscale the global climate model outputs to the flood inundation simulations in Jakarta. The analysis indicated that the combination of climate change and urban development amplified the mean future flood risk in Jakarta. The results also showed large uncertainty for the future flood risk from future urban-growth and climate-change scenarios. Therefore, we conclude that a flood mitigation plan should consider land subsidence, land-use change, and climate-change.

In Japan, typhoons and frontal rains cause severe water-related disasters almost annually, resulting in considerable damage to human life and property. Although multiple hazard and risk evaluations have been conducted in Japanese rivers, hazard evaluations of smaller rivers and tributaries managed by prefectures are unsatisfactory compared with those of the larger rivers managed by the national government. Several flood-related disasters occur in these small rivers because of insufficient data and risk analysis. In addition, the river hazard map issued by the government is a flood inundation diagram that has caused multiple embankment failure, and it is unknown where the most dangerous. In this study, rainfall-runoff simulation was conducted on all rivers in Toyama Prefecture to evaluate flood hazard areas in the prefecture by evaluating their overflow, erosion and seepage potentials. The results of the evaluation of each potential indicate the risk of overtopping due to water level rise at the confluence and the risk of erosion at the bend. Erosion was highly rated in the overall evaluation, and the estimation of risk areas reflects the erosive effect of rapids, which is a characteristic of rivers in Toyama Prefecture.  Also, regional climate change impacts on the flood risks and disaster potentials are analyzed based on the model and d4PDF dataset, and then a vegetation management in the rivers are proposed as an adaptation measure to the climate change.

In recent years, the intensity and frequency of floods have increased in Japan because of its geological characteristics with two-thirds of its land covered by forests and climate change. The Naruse River basin suffered severe damage by Typhoon 19^th in 2019. To mitigate such serious flooding, “Integrated Flood Control in Basin,” which utilizes various potential approaches in the whole basin, is now under consideration in Japan. One of the approaches is the paddy field dam with low cost and high effect. It takes advantage of the paddy fields’ temporal storage capacity by installing drainage control devices at drain outlets. Since 21 % of the land in the Naruse River basin is paddy fields, the potential water storage of paddy field dams is high. Thus, the objective of this research is firstly to develop a simple paddy field dam model that can be applied to the whole basin, and secondly to estimate the potential flood mitigation effect of the paddy field dam under Typhoon 19^th in 2019. Finally, to evaluate the distribution of flood mitigation effect of paddy field dam under the designed rainfall and estimate actual water storage in the entire basin and peak cut distribution using paddy field dams. To approximately estimate the potential effect of different water storage in the paddy field dam applied in this whole basin, scenarios for changing the height of free-drain and application rate of the paddy field dam were evaluated. In conclusion, the paddy field dam has a significant effect on flood mitigation under Typhoon 19^th in 2019. The actual water storage of the entire basin is 70 million m³ under the design rainfall using the 15 cm height of free-drain and the effect of peak discharge reduction would be obvious if the paddy field dams were applied in the upstream reach.

 Serious flood disasters have occurred frequently in the world due to extreme heavy rain and huge flood more than design flood. It is important to clarify the function of flood-control measures against large scale floods more than design flood (hereafter as large flood). This study aims to examine flood-control effect of two retarding basins with different height of overflow weir during large floods. The target sites chosen in this study were the Arakawa River 1st retarding basin and the Ichinomiya 2nd retarding basin in Japan, in which huge floods were caused by Typhoon No.19 and No.21 in 2019, respectively. Ichinomiya River 2nd retarding basin is located middle reach of the Ichinomiya River. Its planned storage is 0.70 million m³. In Typhoon No.21, 6-hours basin-averaged rainfall exceeded the design rainfall by 1.5 times, and finally it was filled, and overflowed. To clarify flood control effect of it, we also conducted flood simulation with 2D horizontal numerical model. In the simulation, we changed the height of overflow weir of it. The results showed the limitation of flood-control function of the retarding basin under huge flood like Typhoon No.21 even though the height of overflow weir were changed. The Arakawa River 1st retarding basin is located middle reach of Arakawa River. In Typhoon No.19, it stored about 35 million m³. In this simulation, we gave two conditions for river discharge in which one is the control condition and the other is the larger flood. The results showed that the retarding basin can maintain the flood control function even for the large flood due to Typhoon No.19. It is also noted that selecting appropriate heights of the overflow weir is important to reduce flood discharge and water level in the downstream reach of the rivers.

Precipitation increases frequently due to climate change. This effect has already been recognized as a phenomenon, and the damage caused by constant floods is increasing. Against this background, urgent infrastructure development is needed. However, the budget is imminent, and new and robust measures cannot be developed. As a response to climate change, approaches to improve existing land use and facilities to store floodwaters should also be considered. In this study, we extracted landuse and facilities that could store water in the Japanese archipelago(landuse and facilities: Puddy field, school area and park). In addition, the potential of this water storage was clarified as spatial information.
As a result, we were able to estimate the amount of possible reduction of inland waters measures in Japan.

Accurate knowledge of the long-term precipitation information is crucial for understanding the mechanisms of the precipitation in coupling Earth’s water fluxes, energy balances, and biogeochemical cycles across space-time scales under the changing climate. And the emerging of ERA5-Land (0.1°, hourly, 1950-present), from the fifth generation of ECMWF, will surely provide great opportunities for exploring the much greater details in the long-term evolution of the precipitation events. However, it is inherently affected by non-zero, and often significant random errors and biases. Base on the characteristics of the high spatiotemporal resolutions and continuity of ERA5-Land, and the fine quality of the APHRODITE, we have proposed a new approach (Daily Total Volume Controlled Merging and Disaggregation Algorithm, DTVCMDA) for generating a new long-term precipitation data, AERA5-Asia (0.1°, 1 hourly, 1981-2015, Asia), and soon the AERA5-Asia back extensions (0.1°, 1 hourly, 1951-1980, Asia) for the Asian precipitation-related scientific research and societal applications. The main conclusions include but are not limited to the following: (1) the quality of ERA5-Land approximates that of IMERG-Final but with larger biases; (2) APHRODITE are firstly and reasonably updated over the no data grids through a novel scheme; (3) AERA5-Asia significantly outperforms ERA5-Land in terms of magnitudes and precipitation occurrences, and is to be on a par with AIMERG. Additionally, results of this study suggest that it would provide a valuable reference in anchoring schemes using gauge analysis for generating the future GPM Level 4 global precipitation products, targeting on its potential foreseeable drawbacks.

Quantification of irrigation water use (IWU) is crucial to understanding the anthropogenic disturbance on  hydrological cycle and is also beneficial to optimal agricultural water management. However, the conventional survey-based method is difficult to obtain the time series IWU at a large scale, while the currently existing satellite-based IWU estimates are subjected to uncertainties related to data gaps and model physics. To overcome the deficiencies, we proposed a framework to couple processes associated with irrigation and integrate multiple satellite-based data to estimate the global IWU. Specifically, the ensemble IWU (IWU_Ensemble) was derived by the balanced relationship of soil water dynamics via integrating a total of 32 combinations of satellite-based soil moisture and precipitation products. The PT-JPL model was incorporated to better characterize the crop evapotranspiration and the DE-MC scheme was employed to improve the efficacy in balancing the soil water dynamics. The IWU_Ensemble demonstrated improved performance than the IWU obtained from individual satellite observations, which are comparable to the conventionally reported irrigation water withdrawal at states of USA (Bias=-0.42km³), provinces of China (Bias=-3.10km³), and country statistics of FAO (Bias=-10.84km³). Larger amounts of IWU were found in India, China, USA, Europe and Pakistan that together contribute to more than 70% of global IWU. A general underestimation of IWU was found in this work and previous studies due to different definitions of IWU and IWW, coarse spatial resolution and asynchronism of satellite observations, use of a static map of irrigated areas, and deficiency in detecting irrigation events in irrigated areas with saturated soil water. Nevertheless, our study demonstrated advantages in integrating multiple satellite observations to reduce uncertainties in estimating the global IWU, manifesting in the improvement of IWU estimates compared to previous studies. Additional work is still needed to produce high-quality satellite-based products for further improving the accuracy of IWU.

Glaciers are the major water source of Lhasa river and Nam Co lake in the Western Nyainqentanglha Mountains (WNM) and play an important role in ecosystem stability and food security of this region. According to previous studies, glacier was continuously shrinking after 2000 and experienced accelerated mass loss in recent years. Glacier surface albedo (hereafter referred to as the albedo) is one of the most important parameters to determine the net shortwave radiation and therefore affect glacier energy and mass balance, thus can help us better explore the causality of glacier mass balance change. With more than 20 years observations, Moderate Resolution Imaging Spectroradiometer (MODIS) is very good data to explore long-term variability of albedo over large regions. Using our new developed albedo retrieval method on snow-ice surface based on MODIS reflectance observations, we analyzed the spatio-temporal variability of glacier albedo in the WNM region during 2001-2020. The results showed that: 1) the mean albedo of the WNM glaciers was 0.552 and the albedo experienced large inter-annual variation and a decreasing trend with 0.0044 a^-1 in the past two decades; 2) all seasonal albedos were decreased and summer albedo decreased fastest, followed by autumn and winter albedos, spring albedo decreased slowest; 3) different glacier albedo changes were observed between north and south slope of the WNM in summer and winter, i.e. the albedo on the north slope was lower but decreased faster than that on the south slope in summer, while the albedo change in winter is the other way round; 4) precipitation and temperature showed high correlation with albedo and could be the main two driving factors for albedo variation; 5) there was a good agreement between albedo and mass balance in terms of spatial variability, which indicate albedo is an important driving factors for mass balance.

Floods are one of the most dangerous and affecting climate-related disasters, and climate change alters the intensity and frequency of floods worldwide. This study examined long-term changes in flood characteristics (including magnitude, frequency and timing) in 30 typical alpine headwaters in the large endorheic Tarim River Basin, central Asia. The contributions of climatic factors to flood changes were investigated using numerical experiments and random forest approach. The results indicate that: (1) annual maximum flood peaks in the area increased at most stations during 1960-2015, with increased flood frequency. Earlier flood peaks were observed in spring, advancing at a rate of 1.38 day/10a, while for other seasons, changes in the occurrence time of flood peaks showed strong spatial variability. (2) In these alpine catchments, catchment characteristics such as runoff coefficient, runoff depth, glacierization and slope had great impacts on flood magnitude, while glacierization and channel slope affected flood timing. (3) Climatic factors including the 3-day antecedent precipitation (P3), 7-day antecedent temperature (T7) were the most influential meteorological factors to flood magnitude. Precipitation was the dominant factor leading to the increase of flood magnitude in most catchments of the south slope of the Tianshan Mountains, while temperature played a greater role in the northern Kunlun Mountains. For flood timing, MLH was the most important influential factor in the alpine catchments in central Asia.

Asian summer monsoon circulation is characterized by strong atmospheric heating over rather high latitude regions. The heating process is separated into latent heating due to the convective activities and sensible heating associated with the highly elevated Tibetan Plateau. To elucidate the impact of the Tibetan Plateau, the heating process of Asian summer monsoon was investigated using recent reanalysis ERA5 with high spatio-temporal resolution. Based on the mass conservation law of upper tropospheric high dry static energy airmass (HDSEA, defined as the airmass with potential temperature from 360K to 375K), the mass source due to latent and sensible heating and sink due to the radiative process were evaluated. The source of the HDSEA should be high moist static energy airmass (HMSEA, defined as the lower tropospheric airmass with equivalent potential temperature over 360K), which is produced at the boundary layer. The amount of HMSEA in the lower troposphere was evaluated to find the mass source of HDSEA. The HMSEA was confirmed to be highly unique to the Asian summer monsoon region. The mass source of HDSEA over the Tibetan Plateau was found in July and August. Highly elevated ground surface produces high potential temperature airmass over the plateau, which partly contribute to the Asian monsoon ciruculateion. However, during the Asian monsoon onset phase, the HDSEA source was not found over the Tibentan Plateau but over the region over the Bay of Bengal and the Bengal Plain. The HMSEA was produced especially over the Bay of Bengal, and was transported to the Bengal Plain. The convective activity over this region played important role in the Asian monsoon onset. Distribution of HMSEA changes as the seasonal march. Associated land surface and convective processes that controls the HMSEA production in the boundary layer will be discussed.

The exchange of heat and water vapor between land surface and atmosphere over the Third Pole region plays an important role in Asian monsoon, westerlies and the northern hemisphere weather and climate systems. A Third Pole Environment (TPE) observation and research Platform (TPEORP) is now implementing over the Third Pole region. The background of the establishment of the TPEORP, the establishing and monitoring plan of long-term scale (5-10 years) of it will be shown firstly. Then the preliminary observational analysis results, such as the characteristics of land surface energy fluxes partitioning and the turbulent characteristics will also been shown in this study. Then, the parameterization methodology has been proposed and tested for deriving regional distribution of net radiation flux, soil heat flux, sensible heat flux and latent heat flux (evapotranspiration (ET)) and their variation trends over the heterogeneous landscape of the Tibetan Plateau (TP) area. To validate the proposed methodology, the ground measured fluxes of the TPEORP are compared to the derived values. The results showed that the derived land surface heat fluxes over the study areas are in good accordance with the land surface status. These parameters show a wide range due to the strong contrast of surface feature. And the estimated land surface heat fluxes are in good agreement with ground measurements, and all the absolute percent difference in less than 10% in the validation sites. The sensible heat flux has increased slightly and the latent heat flux has decreased from 2001 to 2018 over the TP. It is therefore conclude that the proposed methodology is successful for the retrieval of land surface heat fluxes and ET over heterogeneous landscape of the TP area. Further improvement of the methodology and its applying field over the whole Third Pole region and Pan-Third Pole region were also discussed.

High-accuracy meteorological datasets are urgently required for understanding hydrological processes across the Third Pole (Qinghai-Tibetan Plateau, or TP), where meteorological stations are sparse. Low-resolution weather and climate simulations have significant errors in this region due to their inability to resolve meso-micro scale processes associated with the complex terrain and convective clouds. This work presents a one-year dynamical downscaling of the ERA5 data at very high-resolution (approximately 3km) by WRF model. Results show that the kilometer-scale horizontal grid spacing simulation (WRF3) outperforms the ERA5 and the High Asia Refined regional reanalysis (HAR10), in terms of smaller biases and root mean square errors, as well as higher spatial pattern correlation coefficients for 10-m wind speed and precipitation. However, such high-resolution numerical simulation is difficult to be used to obtain long-term precipitation datasets because it is computing resources consuming, while reanalysis data has a long time-coverage and can provide reasonable large-scale spatial and temporal variability. Thus, we further present a statistical downscaling approach to obtain long-term high-resolution precipitation dataset by combining the ERA5 reanalysis with this short-term high-resolution numerical simulation. The approach consists of two main steps: first, the ERA5 precipitation was corrected by the high-resolution numerical simulation at coarse spatial resolution at annual scale; second, the corrected data was downscaled using a convolution neural network (CNN) based model at daily scale. Results show that this approach can reproduce the spatial characteristics of precipitation from the high-resolution numerical simulation, which has a finer spatial structure than ERA5. Evaluation based on rain gauge data shows that the downscaled precipitation has remarkably lower biases than that in ERA5 over the TP. Moreover, it even has higher accuracy than the high-resolution simulation data with smaller bias and RMSE.