ABSTRACT. Intensifying rainfall and sea level rise under climate change are increasing the frequency and severity of urban floods, making timely evacuation planning critical. To support evacuation decision-making under urban flooding, this study develops a multi-objective phase-adaptive evacuation optimisation model. The road system is represented as a directed graph, and a network flow analysis is applied to jointly account for link capacities, travel distances, intersection nodes and shelter capacities, whilst explicitly incorporating flood risk on road segments located in potentially affected areas. Hydraulic simulations are first conducted to characterise spatiotemporal flood dynamics; flood arrival times, water depths and inundation extents are then translated into risk scores and safe time windows for each road segment and neighbourhood. These hazard variables serve as key inputs to the optimisation for evacuation risk assessment. The evacuation problem is formulated as a minimum-cost network flow model, in which routes to shelters are represented as a directed network of nodes and arcs with specified capacities and cost per unit of flow on the arcs. The multi-objective optimisation seeks to minimise (i) total evacuation completion time, (ii) cumulative risk exposure along routes, and (iii) imbalance in shelter utilisation, whilst generating recommended routes from evacuees’ origins to shelters. An agent-based model (ABM) is then employed to simulate individual movements during evacuation for verification. Because ABMs capture the actions, reactions, and interactions of autonomous individuals within a complex environment, we implement the evacuation simulations in NetLogo to reflect dynamic behavioural responses during flood events. Discrepancies between the theoretical optimum and ABM outcomes are evaluated across multiple scenarios and fed back to iteratively refine model assumptions and parameters. The study ultimately delivers optimised evacuation networks, identifies critical bottlenecks, and proposes strategy improvements to support local governments in evacuation planning and emergency response under urban flooding.

ABSTRACT. Heavy rainfall events have become increasingly intense and frequent due to factors such as climate change, resulting in severe flood damage across Japan. To mitigate such impacts, previous studies have developed systems for real-time water-level monitoring and visualization at multiple locations, as well as short-term water-level forecasting using numerical simulations. However, most existing approaches focus on Class A and Class B rivers and are not designed for rural area. Furthermore, these systems generally lack real-time visualization of inundation conditions when water levels exceed drainage canal wall heights. To address these limitations, we developed a cloud-based application specifically targeting rural regions characterized by agricultural land use and small-scale drainage canals. The application provides real-time visualization of canal water levels monitored by ultrasonic sensors equipped with communication capabilities. In addition, it enables hydraulic simulation using a one-dimensional unsteady flow model that incorporates data assimilation techniques. The model physically represents both runoff from adjacent farmland to drainage canals and overflow from canals to farmland, which are critical for reproducing inundation phenomena. Data assimilation scheme, extended Kalman filter, allows continuous integration of recent water-level observations, while forecast rainfall data are used to compute water levels up to 12 hours ahead at 20-minute intervals. Both monitoring and prediction results are rendered in three dimensions through a web-browser-controlled application running on the cloud. The system reconstructs canals and surrounding farmland in 3D, displaying water level and simulating inundation when canal walls are overtopped. The application is currently being trialed by water facility managers for operational use.

ABSTRACT. Urban pluvial flood simulation strongly depends on the availability of detailed sewer network data, including pipe diameters, invert elevations, and network topology. However, such information is often unavailable, incomplete, or difficult to obtain for many cities, which limits the practical applicability of physically based urban flood models. This study proposes a simple and transferable framework to construct a virtual sewer network solely from open road network data and to enable pluvial flood analysis without relying on existing sewer datasets. The proposed framework generates a virtual sewer system by placing sewer pipes along road centerlines and defining network topology using graph-based rules. Pipe diameters and invert elevations are not hydraulically designed in a conventional sense but are assigned using simplified, distance-based criteria with respect to drainage outlets such as pumping stations. This approach allows sewer attributes to reflect large-scale drainage hierarchy while avoiding detailed design calculations. The method is computationally light, scalable to large urban areas, and readily applicable to cities with multiple drainage outlets. The virtual sewer networks are coupled with a hydrodynamic 1D–2D urban flood model using a slot-based sewer representation, enabling simultaneous evaluation of surface inundation depth and flood volume. The proposed approach is applied to Osaka City, Japan, where comprehensive sewer network data are available. Simulation results obtained using the virtual sewer networks are systematically compared with those using the actual sewer system. The results indicate that the adopted strategy for constructing virtual sewer networks strongly affects simulated inundation depth and flood volume. In particular, sewer networks generated using a minimum spanning tree (MST)-based topology exhibit improved agreement with simulations based on the actual sewer network, especially in terms of spatial patterns of inundation depth and total flood volume. These results highlight the importance of appropriate network representation when simplified sewer attributes are used. The proposed framework enables rapid construction of plausible sewer networks for pluvial flood assessment in data-scarce urban areas. In addition to standalone applications, future work will explore hybrid use of the proposed virtual networks in combination with partially available or incomplete sewer datasets, allowing complementary enhancement of existing sewer information for urban flood analysis and drainage planning.

ABSTRACT. Flooding is one of the most destructive natural hazards worldwide, causing severe human casualties and substantial economic losses, particularly in rapidly urbanizing coastal cities. Accurate flood susceptibility mapping (FSM) is therefore essential for disaster risk reduction, land-use planning, and sustainable urban development. This study aims to develop a comprehensive GIS-based FSM framework for Da Nang City, Vietnam, by integrating multi-source geospatial data with advanced machine learning (ML) techniques. Four widely used ML algorithms—Logistic Regression (LR), Support Vector Machine (SVM), Neural Network (NN), and an ensemble model based on K-Nearest Neighbors (KNN)—were implemented and comparatively evaluated. Ten flood conditioning factors were selected based on hydrological relevance and data availability, including topographic variables (elevation, slope, and aspect), land use/land cover (LULC), rainfall, distance to river, and four remote sensing–derived indices: Topographic Wetness Index (TWI), Stream Power Index (SPI), Normalized Difference Vegetation Index (NDVI), and Normalized Difference Water Index (NDWI). Historical flood inventory data were compiled from documented inundation extents associated with the major flood event that occurred in Da Nang City in 2022, and combined with non-flood samples to construct the modeling dataset. The dataset was randomly divided into 70% for training and 30% for independent testing. To enhance model generalization and minimize overfitting, a k-fold cross-validation strategy was employed during the training process. Model performance was assessed using classification accuracy and the area under the receiver operating characteristic curve (AUC). The results indicate that all four models achieved strong predictive performance, with accuracy values exceeding 90% and AUC values ranging from 0.944 to 0.9898. Among the tested models, the SVM demonstrated the highest predictive capability, followed by NN, the ensemble KNN model, and LR. The resulting flood susceptibility maps show consistent spatial patterns, with high-risk zones predominantly located in low-lying and densely urbanized areas. Feature importance analysis using the Chi-square test indicates that rainfall, LULC, and NDVI are the most influential factors governing flood occurrence, underscoring the combined effects of extreme precipitation and urban surface characteristics. Overall, the generated FSMs provide valuable spatial insights for flood risk management, early warning system development, and urban planning in flood-prone coastal cities. The proposed GIS–ML framework is transferable and can be readily applied to other urban regions with similar hydro-climatic and land-use conditions.

ABSTRACT. High-resolution urban flood modelling is essential for effective urban runoff and flood management, particularly for disaster mitigation and early warning applications. However, the use of meter-scale grids over large urban domains leads to substantial computational demands, which often limit the applicability of high-resolution models in operational forecasting and scenario analysis. To overcome these limitations, this study explores the use of multi-GPU acceleration to enhance the computational efficiency of large-scale urban flood simulations. We develop a multi-GPU version of the H12 two-dimensional urban flood model using an MPI–OpenACC hybrid framework. The H12 2D model is a physics-based surface flow model originally designed for CPU-based parallel computation and is extended in this study to efficiently utilize GPU architectures. The proposed framework adopts directive-based parallelization, allowing a single code base to run on both CPU and GPU systems with minimal code modification. MPI is used for domain decomposition, while computationally intensive kernels are offloaded to GPUs through OpenACC directives. This design ensures portability across heterogeneous high-performance computing environments and enables efficient utilization of multiple GPUs. Model performance is evaluated across spatial resolutions ranging from 1 to 20 m for two contrasting study areas: a highly urbanized catchment in downtown Portland, Oregon (USA), and a downstream reach of the Han River basin (Republic of Korea). The results demonstrate how computational performance and scalability vary with grid resolution, domain size, and GPU workload distribution. The improved computational efficiency achieved through the proposed multi-GPU framework enables pseudo real-time high-resolution urban flood simulations, supporting early warning systems and operational flood management. Furthermore, the framework facilitates large-scale, high-resolution simulations suitable for generating ground-truth datasets for the development and validation of physics-based and data-driven flood prediction models.

ABSTRACT. Abstract

Climate change and rapid urban development are creating new challenges for stormwater management in New Zealand. Traditional drainage systems often struggle with extreme rainfall, resulting in flooding, infrastructure damage, and environmental degradation. This study examines the potential of integrating Artificial Intelligence (AI) and smart sensors with a wetland system to improve stormwater performance. The case study focuses on the 120 Willowbank Avenue development in Napier, using hydrological modelling (HEC-HMS) and conceptual design (Civil 3D) to analyse wetland behaviour under varying rainfall conditions. Findings suggest that AI-assisted wetlands can support real time monitoring, predictive maintenance, and adaptive management, contributing to more resilient and sustainable urban water systems. Recent extreme events such as Cyclone Gabrielle highlight the limitations of New Zealand’s conventional infrastructure, which remains largely reactive. Wetlands effectively treat stormwater and reduce runoff but operate as passive systems. Emerging technologies like AI and smart sensors can enhance these systems by enabling predictive management and data driven decision making. This study investigates how such integration could improve wetland-based stormwater management, using the Willowbank Avenue site as a representative example. A mixed-method approach was used, combining hydrological modelling, conceptual design, and expert feedback. Rainfall and catchment data were obtained from HIRDS and local IDF curves, while NCC and HBRC standards guided the wetland concept. HEC-HMS modelling simulated rainfall–runoff responses and assessed the wetland’s potential to reduce peak discharge and improve detention efficiency. Civil 3D supported the conceptual layout of drainage paths, wetland zones, and sensor placements. A short expert survey involving professional engineers and council officers evaluated feasibility, addressing cost, reliability, maintenance, and benefits. Results suggest that AI and smart sensors can enhance prediction, enable early failure detection, and support proactive maintenance. They may optimise wetland operation and improve flood control and water quality outcomes. Survey participants generally supported AI integration but raised concerns regarding initial cost,

calibration reliability, and maintenance needs. Many noted that AI-based systems are still at an early stage and require pilot testing and field data. A hybrid approach, combining natural wetland treatment with AI-assisted predictive control, was considered the most practical. Overall, this study demonstrates the conceptual feasibility and benefits of integrating AI and smart sensors into wetland-based stormwater management. Although limited to conceptual modelling and expert input, the findings provide a framework for future research and pilot implementation in New Zealand.

ABSTRACT. Dili, the capital of Timor-Leste, for the past few years has encountered floods due to poor drainage systems that frequently fail to manage and convey the surface runoff from the heavy rain. Rapid urbanization and climate change have exacerbated these drainage system failures, resulting in significant economic, social, and environmental consequences. This research aims to assess the performance of the urban drainage system and to conduct a flood analysis in Kampung Alor, Dili, Timor-Leste, by using the PC Storm Water Management Model (PCSWMM) as the primary hydrodynamic modelling tool. The ground-level data, historical rainfall and tidal data recorded on 3-4 April 2021, and the drainage master plan of Dili were incorporated into the model. The dynamic wave routing method was applied to allow ponding in 1D simulation. A total of six drainage lines were examined in 1D simulations. Most of the conduits were operated at full capacity on 4 April 2021, leading to flooding in Kampung Alor and along Maloa River. For the 2D surface component, the Digital Elevation Model was processed and converted into grid points and mesh cells to connect to 1D junction attributes to simulate overland flow pathways. The boundary conditions were assigned using rainfall (upstream conditions) and tidal levels (downstream conditions). The water elevation profiles obtained from 1D PCSWMM for the 4 April 2021 flood event indicated that the existing drainage system is insufficient to manage extreme rainfall and surface runoff effectively, particularly in low-lying areas with inadequate drainage systems. The results show that flooding is mainly caused by limited drainage capacity, flow restrictions, and sudden changes in channel design. Maloa River experienced flooding due to transitions between open and closed sections and bridge obstructions. One drainage line showed flooding due to sharp changes in flow direction causing back flow overall the drainage system in study area released the total flood volume of 1.6x106 m3. This is further confirmed by the flood inundation extents simulated using 2D PCSWMM. This study underlines the importance of reducing flow resistance or back flow for flow direction and cross-sectional dimensions in low-lying areas particularly within drainage system with narrow conduit to guaranty the effectiveness of drainage system in transporting the surface runoff properly to outfalls.

ABSTRACT. Recent climate change has led to a significant increase in extreme weather events, creating a critical demand for advanced forecasting models that ensure urban safety while overcoming the heavy computational burdens of traditional physical-based simulations. Urban areas are particularly vulnerable to compound flooding—a complex phenomenon where elevated river levels hinder drainage, resulting in simultaneous sewer surcharges across complex urban geometries. While multi-dimensional hydraulic models offer high accuracy, their substantial processing time makes them impractical for real-time early warning systems or extensive uncertainty analysis involving numerous scenarios. To address these challenges, this study proposes a deep learning-based surrogate modeling framework designed to provide high-resolution spatial and temporal predictions of grid-level inundation depths in real-time.

The Naengcheon Basin in Pohang, South Korea, was selected as the primary study area due to its structural vulnerability to both external and internal water flooding. A high-fidelity training dataset was generated through a physics-based 1D-2D coupled numerical simulation framework. This approach integrated sewer network flow with surface runoff to accurately represent the interaction between the drainage system and overland flooding across 144 diverse scenarios, encompassing return periods from 10 to 500 years. The proposed model integrates Graph Convolutional Networks (GCN) to capture spatial features from unstructured computational grids with Gated Recurrent Units (GRU) to learn the dynamic temporal patterns of flooding. By applying a sliding window technique to the rainfall and hydraulic input data, the framework predicts continuous inundation sequences at each node.

Analysis results indicate that the GCN-GRU model significantly outperforms baseline models, reducing RMSE and MAE by approximately 14% and 11%, respectively. The framework demonstrated robust predictive capability even under extreme conditions, achieving a Critical Success Index (CSI) of 0.698 and an F1-score of 0.821 for major rainfall events. This research confirms that graph-based surrogate models can accurately replicate complex hydraulic behaviors with high efficiency, offering a scalable and practical solution for rapid flood impact assessment and urban resilience management without the need for repetitive and costly numerical simulations.

Acknowledgement This work was supported by Electronics and Telecommunications Research Institute(ETRI) grant funded by the Korean government [25ZR1300, Development of Technology for the Urban Extreme Rainfall Response Platform].

ABSTRACT. Turbidity events in reservoirs frequently occur following intense rainfall, as sediments are transported from upstream. These events pose critical management challenges due to their impacts on water quality degradation, intake stability, and the overall efficiency of water resource operations. In stratified reservoir environments, where stratification and complex flow dynamics coexist, a detailed understanding of turbidity generation and transport mechanisms necessitates high-resolution, field-based observations. In this study, we developed and applied a multi-sensor turbidity monitoring system tailored for complex reservoir conditions. To overcome the limitations of conventional monitoring methods—such as fixed-point or single-sensor systems—in capturing the spatial and temporal variability of turbid layers, we integrated multiple sensors, including a Conductivity-Temperature-Depth profiler, submersible laser-diffraction based particle size analyzer (LISST-200X), and an acoustic Doppler current profiler (ADCP), into a mobile observation platform comprising both manned and unmanned boats. This system enables vertical profiling at multiple key locations while in motion, facilitating three-dimensional acquisition of stratification structure, suspended sediment concentration and particle size distribution, and flow velocity data. Field observations were conducted four times in Lake Soyang between 2024 and 2025 using the proposed system. Following heavy rainfall, a thermocline was consistently observed at depths of approximately 10–40 m, serving as a boundary below which the turbid layer was predominantly confined. At the thermocline and turbidity boundary, relatively coarser particles were dominant while finer particles were more prevalent within the interior of the turbid layer. Downstream flow was dominant within the turbidity layer, indicating that both the retention and transport of turbidity are closely governed by the reservoir’s stratified structure. The proposed multi-sensor monitoring system demonstrated its effectiveness in resolving the three-dimensional behavior of turbid flows in a reservoir environment. This suggests strong potential for future use as an observation framework for advanced hydraulic and hydrological analyses and for improving turbidity management strategies in reservoir systems.

Acknowledgements This work was supported by Korea Environment Industry & Technology Institute (KEITI) through Research and Development on the Technology for Securing the Water Resources Stability in Response to Future Change, funded by Korea Ministry of Climate, Energy, Environment (MCEE) (RS-2024-00397970)

ABSTRACT. We present Sat2SSC, a novel, low-complexity framework for analyzing suspended sediment concentration (SSC) in large rivers across the Conterminous United States (CONUS) using satellite-based remote sensing and machine learning techniques. We evaluated tree-based regression models for three satellite platforms—Harmonized Landsat Sentinel-2 (HLS), Landsat 8/9, and Sentinel-2—using extensive matchup databases compiled from the Water Quality Portal for period 2015 - 2023. The matchup datasets include 2,540 samples for HLS, 1,103 for Landsat, and 2,707 for Sentinel-2, representing a wide range of riverine conditions. The models were trained using ~60 spectral features, including several band ratios, normalized difference indices, and three-band combinations. Sentinel-2 achieved the highest coefficient of determination (R²=0.642), while HLS exhibited the lowest relative error (0.264). Landsat showed moderate performance, with an R² of 0.518 and a relative error of 0.278. Overall, these results show an added 20 -30% improvement in relative error compared to published benchmarks (0.33-0.39). The framework has successfully been implemented as an interactive Google Earth Engine, as an operational tool, that can inform SSC, support sediment monitoring, provide temporal analysis, and conduct multi-sensor comparisons. The latter of the presentation will be demonstration of the tool, followed up a quick hands-on

ABSTRACT. With advances in image-based flood discharge observation techniques, recent studies have attempted not only to estimate flow velocity and direction on the water surface but also to estimate local water depth. However, most existing studies focus on estimating either the flow velocity and direction or the water depth, and these estimations have generally been discussed separately. Moreover, previous studies by the authors have revealed that the wavenumber–frequency spectrum of water-surface variations contains components due to both advection and wave motion. It is expected that the flow velocity and direction, as well as the water depth, can be estimated simultaneously from the correspondence between these components and the dispersion relations of turbulence and surface waves. This study aims to investigate a method for simultaneously estimating the flow velocity, flow direction, and water depth from a three-dimensional wavenumber–frequency spectrum of water-surface variations using UAV-acquired video images of a river surface. Image analysis was conducted on video images of the river surface taken from above and upstream of Nekoya Bridge on the Uono River in Japan. Square interrogation areas were arranged continuously across the river surface in the transverse direction, with no gaps or overlaps. A three-dimensional wavenumber–frequency spectrum, composed of streamwise and transverse wavenumbers and frequencies, was created by applying a discrete Fourier transform to the extracted time series of brightness variations for each interrogation area. The flow velocity, flow direction, and water depth were simultaneously estimated by matching the theoretical dispersion relations of turbulence and surface waves to the spectral peaks using the Normalized Scalar Product. The estimated flow velocity and direction were in good agreement with those obtained using Particle Image Velocimetry. In addition, the estimated water depth was also generally consistent with the mean water depth at the study site. These results demonstrate the potential for simultaneously estimating flow velocity, flow direction, and water depth from video images of the river surface.

ABSTRACT. Satellite-derived groundwater storage (GWS) products are increasingly used to investigate groundwater variability at regional to global scales. However, a key open question remains: which components of observed groundwater dynamics are represented by these products, and how does this representation vary across aquifer systems? Addressing this question is essential for the appropriate interpretation and application of satellite-based groundwater information. In this study, we develop a process-oriented framework to compare satellite-derived GWS products with dense in situ groundwater level observations from the Poyang Lake Basin, China, a large lake-dominated system characterized by strong surface water–groundwater interactions. The analysis is designed to examine the consistency between observed groundwater levels and satellite-derived GWS across temporal components, aquifer types, and spatial settings. This study systematically explores how different satellite products respond to seasonal forcing, long-term change, and short-term variability in groundwater systems. Rather than aiming to validate or replace in situ observations, this work seeks to clarify the interpretability and limitations of satellite-derived GWS products. The proposed framework provides a transferable approach for understanding what aspects of groundwater dynamics satellite products can—and cannot—capture, offering guidance for their use in groundwater drought analysis and large-scale water resources assessment.

ABSTRACT. FUNWAVE-TVD is a public-domain, fully nonlinear Boussinesq wave model that is widely used in nearshore and coastal applications. It employs a Godunov-type Riemann solver with a total variation diminishing (TVD) scheme and is formulated in conservative form to enhance numerical stability and computational efficiency, while accounting for vertical vorticity effects and a time-dependent reference elevation. In this study, the model was used to investigate the nonlinear evolution of random wave propagation and dissipation over varying depth. The simulations were based on laboratory wave-flume experiments conducted under a range of water depth, wave height, and wave period conditions. The water depth varied from 0.7–1.0 m near the wavemaker to a shallow region of 0.35–0.65 m along the slope. A total of 65 experimental cases, generated using a JONSWAP spectrum with significant wave heights between 0.03 and 0.26 m and peak periods from 1.07 to 6.15 s, were simulated to rigorously validate the model across a broad range of wave conditions. The numerical setup, including grid spacing and the number of wave components, was systematically optimized to accurately reproduce power spectra and key statistical wave parameters, such as significant wave height and higher-order moments (skewness and asymmetry), which characterize the vertical and horizontal deformation of waveforms induced by higher-order harmonics. As a result, optimal numerical configurations were identified that minimize numerical prediction errors while maintaining physical consistency with the experimental measurements.

ABSTRACT. A decade-long global L-band (1.4 GHz) microwave brightness temperature (TB) observations from the Soil Moisture Active Passive (SMAP) mission have provided surface soil moisture (SM) for drought assessment, water resources management, and agricultural productivity. However, their coarse resolution (tens of km) limit to resolving heterogeneous landscapes such as smallholder agriculture and mixed land-cover mosaics, where sub-footprint variability can dominate retrieval uncertainty. High-resolution (up to 200 m) SM products from the NASA-ISRO Synthetic Aperture Radar (NISAR) mission improve spatial detail, yet field-scale information is still needed to characterize within-field SM variability and to support product NISAR SM evaluation.

This study presents a multi-sensor Unmanned Aerial Vehicle (UAV)-based SM retrieval. UAV observations include L-band TB measurements, vegetation indices and thermal infrared temperature from a multispectral sensor, and high-resolution LiDAR. This helps minimize parameter uncertainty in the SM retrieval algorithm. The multi-sensor UAV campaign is conducted over a densely instrumented SM validation site in Hampyeong-gun, Republic of Korea (35.0103°-35.0198° N, 126.5456°-126.5560° E), where in situ SM and soil temperature sensors at 5 cm depth span agricultural, forest, and bare land surfaces.

The retrieved SM data are further downscaled with physically meaningful predictors (ground temperature, vegetation indices, LiDAR elevation). Downscaling is performed with supervised learning algorithm, where UAV-retrieved SM at 30-m serves as the target variable and UAV-derived predictors resampled to 30-m for training. The trained model is applied to obtain cm-scale SM that resolve within-field variability. Agreement and scale-dependent uncertainty are quantified using correlation coefficient, ubRMSD, and bias.

This study is expected to not only provide an initial blueprint of NISAR SM retrieval accuracy across various land cover types but also an understanding of cm-scale high-resolution SM variability from UAV-based downscaled SM data.

ABSTRACT. Lake Paldang is an artificial reservoir formed by a dam at the confluence of the Namhan River, Bukhan River, and Gyeongan Stream in Korea. Owing to its location within a large watershed, the lake is characterized by a short hydraulic retention time and strong riverine influence, resulting in pronounced spatial variability in hydrodynamic behavior. Distinct flow and mixing characteristics emerge across the riverine, transitional, and lacustrine zones, reflecting the combined effects of inflow conditions and hydraulic regulation. In particular, the Namhan and Bukhan Rivers are regulated by multiple upstream dams, causing substantial temporal variations in discharge and water temperature. These variations strongly influence density-driven inflows, stratification development, and internal circulation within the reservoir.

As a river-dominated lake, Lake Paldang is hydraulically sensitive to changes in inflow conditions and structurally vulnerable to water-quality degradation. Consequently, understanding three-dimensional flow and mixing processes originating from river confluences is essential for accurately interpreting spatial and temporal variations in water temperature, flow structure, and water-quality dynamics. However, such complex hydrodynamic behavior cannot be sufficiently described using conventional depth-averaged or point-based observations alone.

This study aims to analyze the hydrodynamic characteristics of river flow and mixing behavior in Lake Paldang using spatially resolved field measurements. High-resolution cross-sectional and vertical profiles of flow velocity and water temperature were collected using advanced monitoring equipment, and spatial information analysis was applied to investigate three-dimensional flow structures and mixing patterns. The results indicate that the formation, strength, and spatial extent of density currents vary markedly depending on upstream discharge conditions and temperature differences between inflows. Furthermore, the operation of upstream hydraulic structures exerts a dominant control on internal circulation patterns and mixing efficiency within the reservoir.

The findings demonstrate that hydrodynamic processes in river-dominated lake systems are inherently three-dimensional and highly heterogeneous in space. Spatial information–based hydrodynamic analysis provides critical insights into flow and mixing mechanisms that cannot be captured by simplified approaches. This study underscores the importance of continuous, spatially resolved monitoring for improving the understanding of river flow behavior in regulated reservoirs and offers valuable implications for adaptive reservoir operation and effective water-quality management.

This work was supported by the Ministry of Climate, Energy, Environment through the project ‘Research and Development on Technology for Securing Water Resources Stability in Response to Future Changes (grant number RS-2024-00332114)

ABSTRACT. Understanding stratification and water flow structures in rivers is essential for interpreting hydraulic processes and their influence on water quality and material transport. Rivers have traditionally been assumed to be vertically well mixed due to continuous flow; however, recent studies have reported that transient stratification and complex three-dimensional flow structures can develop under specific hydraulic and thermal conditions. This study aims to analyze stratification phenomena and water flow structures in rivers using high-resolution in-situ monitoring data. In-situ monitoring was conducted by traversing river cross-sections using an acoustic Doppler current profiler (ADCP) and a multiparameter water quality sonde (YSI EXO2). Hydraulic and water-quality data, including flow velocity, water depth, and temperature, were simultaneously collected at the water surface and along the vertical direction, enabling the construction of spatially detailed datasets that characterize the hydraulic and thermal conditions of the river. In addition, river cross-sections were defined, and Richardson number (Ri) values were calculated separately for the left, center, and right portions of each cross-section. The Richardson number is an index calculated based on density gradients derived from water temperature and vertical velocity shear; accordingly, Ri was used to evaluate stratification conditions at each cross-section. The results of the Ri analysis showed distinct variations among the left, center, and right portions of the cross-sections, as well as clear seasonal differences. Furthermore, the spatial distribution of Ri enabled the identification of not only stratification but also water flow structures and flow behavior within the river. Relatively high Ri values were associated with stable flow conditions characterized by limited vertical mixing, whereas low Ri values indicated zones where mixing was enhanced by shear. These results suggest that Ri can be utilized not only as an indicator of stratification but also as a flow-related index for interpreting water flow structures in rivers. Therefore, this study demonstrates that combining high-resolution, simultaneous surface and vertical measurements with laterally resolved Ri analysis provides an effective analytical framework for integrated evaluation of stratification and flow dynamics in river environments.

This work was supported by the Ministry of Climate, Energy, Environment through the project ‘Research and Development on Technology for Securing Water Resources Stability in Response to Future Changes (grant number RS-2024-00332114)

ABSTRACT. Satellite-measured significant wave height is widely used in coastal and ocean engineering fields as an alternative source for ocean wave data when validated with reliable buoy measurements. Satellite-measured significant wave height (Hs_sat) data are commonly employed as fundamental inputs in wave-related research, e.g., numerical model validation and marine structure design. However, Hs_sat data are known to exhibit measurement errors of approximately 10% and therefore require thorough validation against buoy measurements prior to application. However, given the absence of standardized procedures for validating Hs_sat, accuracy assessments and validations are typically conducted using arbitrary approaches and are highly dependent on the applied sampling methods and criteria. The selection of an allowable spatial distance between satellite and buoy measurements is one of the most influential factors in data sampling that affects error metric estimates and Hs_sat validation results. The allowable distance needs to be large enough to include a minimum sample size to ensure robust validation, while small enough to avoid misinterpretation from samples exhibiting large distance. As aforementioned, however, no guidance has been established for selecting the allowable distance yet and most previous studies have arbitrarily defined the allowable spatial distances of 0.25° or 0.5°, introducing substantial variability and uncertainty. In the present study, the effect of selecting the allowable distance between buoy and satellite measurements on Hs_sat validation is systematically investigated using 107 offshore buoy datasets from the U.S. National Data Buoy Center and satellite measurements data from 17 satellites provided by the Australian Ocean Data Network. This effect is quantified by estimating linear trends of error metrics (e.g., correlation coefficient, scatter index, and percent bias) as the allowable distance increases. While error metrics generally increase with increasing allowable distance, their magnitude and patterns vary among buoy locations. The results show that the linear trends in error metric estimates at coastal buoys are relatively steeper than those at offshore buoys. In addition, buoy stations where error metrics remain consistent regardless of the selected allowable distance are identified. Because determining an appropriate allowable distance remains challenging, adopting buoy locations that exhibit low variability in Hs_sat accuracy assessments can help reduce the addressed uncertainty. Although this study focuses on wave databases along the United States, applying the same approach to other regions is expected to reduce the uncertainty in accuracy assessments of global Hs_sat and enhance the reliability of satellite measurements.

ABSTRACT. The scour caused by storm flow poses a threat to the safety of marine structures and ecosystems, but there are few studies on the storm velocity. The wind speeds and pressures during severe typhoon Muifa (2212) and super typhoon Chanthu (2114) are calculated by using four parameterized wind modes, and compared with the measured values. Results indicate that advanced wind mode, taking the sea surface drag into account, has the highest accuracy in simulating the typhoon wind speeds. On the basis of combining advanced wind mode with the concentric circle grid, a new mathematical model for storm tide is established, which offers fast operation speed, smooth grids and local high resolution. The storm surge and velocities in Hangzhou bay and East China Sea were simulated. The storm velocities and water level predicted from the new model agree well with the observed values in Hangzhou Bay during Typhoon 2212. It is demonstrated that the proposed model has high simulation accuracy compared to the traditional models involving rectangular grids with Holland wind mode. Analysis of the storm velocities revealed that surface currents were enhanced significantly during both super typhoon 2114 and severe typhoon 2212. The maximum surface velocity in the Hangzhou Bay reached 6.06 m/s during typhoon 2114, with increment of 1.7m/s compared to that under windless conditions. During typhoon 2212 the maximum surface velocity in East China Sea is 3.81 m/s, 2.89 times of the corresponding velocity at the center of typhoon. Results show that there is an evident plane circulation flow in lower bay owing to the influence of super typhoons. The proposed mathematical model provides an insight into the prediction of typhoon storm tides, which strengthens the scientific foundation for typhoon disaster prevention and reduction, as well as ecological protection efforts.

ABSTRACT. Tsunamis are characterized as low-frequency, high-impact disasters that cause catastrophic damage despite their rare occurrence. Conventional deterministic approaches, which focus on specific worst-case scenarios, have fundamental limitations in quantifying return periods and associated uncertainties. The unexpectedly devastating 2004 Indian Ocean and 2011 Tohoku earthquakes exposed critical gaps in existing hazard assessments, prompting the rise of PTHA as a key analytical tool. The East Sea is a semi-enclosed sea containing the eastern margin fault zone extending from western Hokkaido to the Noto Peninsula, where historical tsunami events have been documented. Furthermore, the eastern coast of Korea, situated within the reach of tsunamis, faces high potential risks due to the concentration of critical infrastructure. Despite these risks, there is still a lack of studies that account for the contributions of various seismic sources through systematic deaggregation for a precise hazard assessment in this region. This study performs a comprehensive PTHA for the entire eastern coast of Korea by integrating six seismic source zones in the East Sea. A logic-tree framework was established to account for epistemic uncertainty, generating 2,160 simulation branches through combinations of variables such as magnitude, fault geometry, dip, and strike. The final scenario set was constructed by incorporating statistical branches, including return periods and uncertainty coefficients. Numerical simulations were conducted using the COMCOT model with a high-resolution 40-meter grid system. The results were interpreted using Poisson-based annual exceedance probability curves and fractile curves to define the confidence intervals of the hazard. Additionally, deaggregation was applied to the PTHA results to quantitatively identify the regional contributions of major seismic sources. The hazard analysis revealed significant regional variations, with Sokcho and Donghae identified as high-risk areas due to predicted significantly high tsunami wave heights. This hazard amplification is attributed to wave energy focusing caused by complex bathymetry, such as the Yamato Rise and the K-shaped ridge. Furthermore, the deaggregation analysis determined that fault ruptures at the central-eastern margin are the dominant factors influencing the hazard in these regions. By systematically isolating the contributions of magnitude and source zones, this study contributes to establishing priority-based disaster prevention strategies grounded in physical interpretation and identifying potential primary risk sources.

ABSTRACT. Wave breaking plays a vital role in enhancing transfer of mass, momentum, and energy across the air-sea interface. Even though wave breaking has been intensively studied, a vast part of the air entrainment process is still poorly understood. One major reason is that multiphase approaches considering air-water mixture density variation have rarely been used. To quantify certain flow properties that involve air entrainment and fluid density variation, such as the mean and turbulent kinetic energy as well as potential energy under breaking waves, void fraction measurements are essential.

In the present study, through combined velocity and void fraction measurements, flow kinematics, turbulence, void fraction, and the effects of void fraction to energy dissipation were investigated. The data uniquely characterize and compare kinematic and dynamic properties of the multiphase flow, including flow structure in the highly aerated region, under plunging and spilling breaking waves.

One of the criteria for predicting breaking onset is through energetic threshold. Barthelemy et al. (2018) proposed a criterion where B is defined as the ratio between the local energy flux and the local energy density projected on the wave propagation direction, and Bth is the threshold value of B at which the wave begins to break. In the present study, B was calculated based on the measured horizontal velocity at the crest peak divided by the phase speed C calculated from the linear wave theory. In the current study, the measured B sharply increases near the breaking onset of Bth = 0.85 under both the plunging and spilling breaking waves. In addition, recorded high-speed video images reveal that a small-scale impinging/splash-up occurred in the process which is likely to cause of the multiple threshold crossings of B.

ABSTRACT. Bormes-les-Mimosas Port is a marina located on the southern Mediterranean coast of France. The port, especially its seawall has experienced gradual deterioration due to strong southerly swells. Severe storm events in 2014 and 2017 resulted in overtopping of the wavewall and breakwaters, threatening berthed vessels and nearby coastal infrastructure. In response, a major restoration project was undertaken to improve the port’s coastal protection performance. Unlike conventional restoration works, the remedial measures adopted must preserve the historical significance and aesthetic appearance of the Memosas Port. In this context, an innovative restoration approach was warranted in place of straightforward classical engineering interventions. Thus, no new structural elements, such as breakwaters or groynes, were introduced. The works rather comprised the raising and strengthening of existing wave wall, reprofiling of embankments, and slope protection using artificial concrete armour units: ACCROPODE and ECOPODE. The ECOPODE units were specifically selected to enhance the visual integration of the structure, as their form and the texture blend harmoniously with the surrounding environment. The coastal protection scheme is represented in the modeling as two cross sections of the revetment: North Profile and South Profile. North profile is expected to experience higher and worse wave conditions than the South profile and thus, the two sections have been designed accordingly. North profile utilizes 6m3 ECOPODE, 8-12T and 6-8T rock armours at toe and back, and a seawall of elevation 5.5m. South profile uses 4m3 ECOPODE, 8-12T and 8-10T rock armours at toe and back, and a seawall with a crest elevation of 4.35m. Salient feature of the two profiles is the channel that runs parallel with the seawall, which collects and conveys a majority of any overtopping discharge. A comprehensive two-dimensional physical modelling program was undertaken at the LHI laboratory to evaluate performance of proposed breakwater profiles under three modes of assessment: structural integrity via a damage level computation, public/structural safety via overtopping discharge computation, and hydraulic pressure estimation on seawall. Testing conditions were based on 10 year, 50 year and 100 year return period waves and corresponding water levels. No armour unit or the wavewall in both profiles moved even under extreme wave and water level condition and thus, none of the armour layers required any further upward modifications in terms of its size nor any stabilization of the seawall. Consequently, the physical model testing provided a high level of confidence in the robustness and adequacy of the proposed design.

ABSTRACT. The environmental impact of dredged sediment disposal in open water is heavily influenced by the initial formation dynamics of the descending sediment cloud. For monodisperse clouds, the formation regime—classified as thermal-like (coherent vortex) or wake-like (elongated, particle-shedding)—has been shown to depend on a dimensionless cloud number, Nc, defined as the ratio of the particle settling velocity to the characteristic circulation velocity within the sediment cloud. However, natural sediments are polydisperse. This study presents a comprehensive experimental investigation into the formation process of composite sediment clouds comprising mixtures of five distinct grain sizes. A total of 27 laboratory experiments were conducted, releasing controlled, quasi-instantaneous sediment masses with varying proportions of coarse to fine particles in a large water tank. The dynamics were analyzed using a composite cloud number, Ncc, which incorporates a buoyancy-weighted settling velocity, thereby accounting for the distribution of grain sizes and their respective contributions to total excess buoyancy. Results demonstrate that the formation regime is effectively governed by the source Ncc. A critical value of Ncc≈3.0×10−2 is identified, separating wake-like behavior (characterized by elongated shapes and significant particle shedding) from thermal-like behavior (featuring coherent vortex structures and minimal sediment loss). Crucially, this critical value aligns closely with that established for monodisperse clouds, indicating a fundamental similarity in the governing physics. The presence of coarser sediments was found to accelerate the disintegration of fine-sediment clumps and stabilize the cloud's vortical structure, promoting integrated descent and reducing ambient turbidity. Quantitative analysis of cloud aspect ratios and frontal descent dynamics confirms that Ncc robustly predicts the formation regime across diverse sediment mixtures. These findings provide a crucial predictive tool for assessing the short-term environmental footprint, particularly turbidity generation, of practical sediment disposal operations involving mixed grain sizes.

ABSTRACT. Global warming accelerates the river inundation and the coastal erosion, so that immediate and effective countermeasures are needed. Sediment flushing from dams and weirs in upstream basins of rivers is an important countermeasure to recover the capacity of them and to maintain coastal beaches. However, the simplistic sediment flushing may lead to the river inundation especially at the top of alluvial fan and to the sedimentation of ship channel at the rivermouth, so that the regional sediment transport model from mountains, river basins to coastal regions is needed in order to conduct the sediment flushing simulations in advance. In this study, the quasi-3D original model named “SPR-WDMPOM” was constructed by taking field data into the model and field applications to the Yamakuni River and Nakatsu Intertidal Flat, which were located in the Kyushu District in the southern part of Japan, were conducted. This model was composed of “RRI model”, which was the rainfall, runoff and inundation model by the ICHARM, the calculation model of river sediment discharge “gRSM” by authors and the quasi-3D coastal sediment transport and morphodynamic model based on the POM “WDM-POM” also by authors. This model took field sediment data into account through “gRSM” in order to improve the quantitative weakness especially on sediment discharges. Numerical results of the flood in 2017 were well agreed with observation results in terms of the hydrograph near the rivermouth, horizontal distributions of mud concentration on the intertidal flat and mud content values at almost all observation points. Observation results also indicated that the siltation of this flat occurred by the heavy flood with landslides of mountains and fine sands were supplied by heavy floods without landslides. Long-term change by the comparison between sediment data obtained in 1984 and those in 2017 revealed that the dominant diameter decreased and the bottom topography became to the mild slope near the east end of this flat. Furthermore, from numerical results of sand capping on this flat, the optimal location and quantity were discussed. Sedimentation of the Yamakuni River Basin was estimated at about 30% of yearly sediment discharge of this river. Sedimentations of the Yabakei Dam and the Heisei Weir were considered as severe problems on the management of river facilities, so that sediment flushing especially from the Heise Weir near the rivermouth was considered as the effective countermeasure not only for the river management but also for the coastal one.

ABSTRACT. This paper conducts an experimental analysis of the nonlinear behavior of normalized bed shear stress induced by tsunami bores traversing smooth, wet-bed surfaces. Prior studies have primarily focused on dry-bed scenarios, leaving the hydrodynamics of wet-bed tsunami bores and their effects on sediment transport and coastal loading insufficiently analyzed. To fill this gap, a series of controlled dam-break experiments was done with a spring-loaded vertical lift gate to create bores at initial reservoir depths of 7 cm, 10 cm, and 15 cm above a constant initial water layer of 5.2 cm. These conditions resulted in dimensionless bore heights of 1.4, 2.0, and 3.0. Using electromagnetic velocity meters and capacitance-type wave gauges, we measured the depth-averaged velocity and the local water depth 5 m downstream. The Saint–Venant momentum equation was used to calculate bed shear stress. Peak velocities increased substantially with bore height, from 0.114 m/s for H = 7 cm to 0.224 m/s and 0.505 m/s for H = 10 cm and 15cm. Likewise, the peak bed shear stress increased, with values of 13.95 Pa, 216.21 Pa, and 242.00 Pa, respectively. However, the normalized peak bed shear stress showed a clear nonlinear, non-monotonic pattern. The highest normalized peak bed shear stress value was at a dimensionless bore height of 2.0, equivalent to 0.640. Smaller and larger bores both produced lower normalized peak bed shear-stress values (at 1.4 equivalent to 0.360, and at 3.0 equivalent to 0.485). This trend shows that wet-bed conditions affect how rapidly the bore loses strength after impact, with energy transfer to the bed being most efficient at moderate bore heights. These findings clarify tsunami-bore seabed interactions and provide helpful information for estimating tsunami intensity, forecasting sediment movement, and designing coastal structures for high-energy tsunami flows.

ABSTRACT. Climate change is increasingly affecting European landscapes through more frequent and severe floods and droughts, accompanied by secondary impacts such as soil erosion, landslides, and wildfire risks. Enhancing climate resilience at the landscape scale has therefore become a central objective across multiple policy domains and is closely linked to the restoration of the soil and water systems. In this context, the European Water Resilience Strategy promotes enhancing the “sponge functioning” of landscapes through the implementation of Nature-based Solutions where possible, complemented with engineered measures where needed.

The purpose of this contribution is to present an integrated scientific and participatory approach that supports the assessment, design, and upscaling of sponge landscape strategies. A key challenge addressed is the lack of methods that can holistically evaluate the performance of individual sponge measures and their combinations under diverse hydrometeorological conditions, while also accounting for ecological and socio-economic impacts.

Building on work from the European projects SpongeScapes and SpongeWorks, we introduce a stress-testing methodology that evaluates current and future sponge functioning at the landscape scale. The approach assesses the effectiveness of measures and strategies across a wide range of flood and drought scenarios and explicitly considers impacts on ecosystem services, natural values, and socio-economic outcomes. Results indicate that individual Nature-based Solutions rarely generate substantial effects at landscape scale on their own; instead, strategically designed combinations of measures are required to achieve robust improvements in water retention and climate resilience.

We further explain how the quantitative results of the stress-testing are integrated in co-creation dialogues using a GeoDesign approach. This approach was tested in the case study ‘Aa-dal Noord’ in the province of Noord-Brabant, The Netherlands. By translating complex hydrological information into spatially explicit and accessible formats, GeoDesign enables stakeholders to explore alternative futures scenarios, compare trade-offs, and jointly develop place-based sponge strategies. Overall, the contribution highlights how combining rigorous assessment methods with participatory design can bridge the gap between science and practice and support the effective upscaling of Nature-based Solutions across European landscapes.

ABSTRACT. Climate change is expected to increase the frequency and intensity of localized heavy rainfall and extreme flooding. Traditional flood defense design methods, based on historical observations, are reaching their limits due to the increasing uncertainty surrounding these extreme hydrological events. Existing river flood management plans primarily focus on installing flood management facilities, such as dams and levees, to efficiently or artificially channel floodwaters. These methods are unsustainable, as they damage the natural integrity of rivers, increase flood risk in downstream areas, and lack the capacity to respond to extreme hydrological events exceeding the design frequency. Securing natural flood buffer zones, such as wetlands, riverside reservoirs, and floodplains, is attracting attention as an alternative to these flood management facilities. However, there is a lack of research and applications for these nature-based flood buffer zone recently. In particular, some existing nature-based technologies focus on water quality and ecological functions rather than flood management. Therefore, this study aims to examine the functions and effects of natural flood buffer zones in areas surrounding rivers, such as existing green spaces, parks, farmland, riverside reservoirs, and floodplains, and, in particular, to discuss a method for quantitatively calculating the flood reduction effects of buffer zones.

ABSTRACT. Floodplain vegetation is a central component of nature-based flood management in river corridors, yet its influence on riparian processes in meandering rivers remains difficult to generalize and scale for practical applications. Vegetation modifies hydraulic resistance, flow partitioning, and channel–floodplain connectivity during floods, with cascading effects on sediment transport, channel morphology, floodplain deposition, and habitat formation. Despite its importance for Nature-Based Solutions (NbS), floodplain vegetation is often simplified in hydraulic models and design guidance, limiting the effectiveness and transferability of NbS strategies across river systems. This work examines how floodplain vegetation density and spatial organization influence hydrodynamic and geomorphic processes in meandering rivers under overbank flow conditions. Key challenges addressed include identifying how vegetation alters flood stage flow structure, determining how these hydraulic changes reorganize sediment pathways and depositional patterns, and assessing how vegetation-mediated processes scale from idealized experiments to natural rivers. Particular attention is given to the role of vegetation in modifying channel–floodplain exchange and secondary circulation, processes that strongly govern flood attenuation and geomorphic response but remain poorly constrained in management-oriented frameworks. This analysis draws on a combination of controlled physical experiments, numerical modeling, and field observations to isolate vegetation effects across a range of flood conditions and spatial scales. Rather than focusing on site-specific outcomes, the emphasis is placed on identifying consistent process relationships linking vegetation density, flow structure, and sediment redistribution. Results show that increasing floodplain vegetation density can substantially reduce channel–floodplain exchange while intensifying in-channel secondary flows, leading to persistent reorganization of shear stress and sediment transport pathways. On floodplains, vegetation promotes spatially heterogeneous deposition and sediment sorting during large floods, reinforcing its role as an active geomorphic agent rather than a passive source of roughness. These findings have important implications for nature-based flood management. Explicitly accounting for vegetation–flow-sediment feedbacks can improve flood-risk mitigation, ecological function, and long-term resilience in meandering rivers. Key challenges to implementation include scaling process understanding across river sizes, representing vegetation dynamics through time, and integrating these effects into flood management frameworks. Addressing these challenges is essential for advancing resilient, adaptive NbS strategies that balance flood protection with ecosystem function.

ABSTRACT. Nature-based flood management increasingly leverages riparian vegetation and floodplain reconnection to attenuate extreme floods while delivering co-benefits such as habitat creation and carbon sequestration. However, implementation and permitting often depend on the credibility of hydraulic predictions, and a key source of uncertainty is how woody riparian vegetation resistance is parameterized and transferred across spatial scales. This study presents an implementation-oriented framework that links full-scale outdoor hydraulic measurements with momentum-based models to produce reliable vegetation resistance inputs for NbS planning and design. The framework integrates three components. First, channel and floodplain geomorphology and existing vegetation conditions are characterized using advanced field surveys and remote-sensing products. Second, resistance for patchy woody vegetation is parameterized using measurable blockage-related descriptors (e.g., effective frontal area and patch arrangement), enabling consistent translation between site observations and model inputs. Third, scenario-based 1D/2D hydrodynamic simulations are conducted to quantify the impacts on water levels, conveyance, and the main-channel–floodplain flow partitioning under alternative vegetation layouts and management options. The central contribution is an evidence pathway that evaluates the reliability of commonly used vegetative resistance formulations under hydraulically complex, compound-channel conditions. Reach-representative measurements from full-scale experiments are used to modify or constrain resistance parameters and to diagnose when direct estimates from plant- or patch-scale drag relationships bias reach-scale friction estimates. The resulting parameter sets and practical guidance support NbS design decisions in river management and engineering, including selecting vegetation configurations that preserve flood-safety margins while maximizing ecological services and enabling the riparian vegetation patches. From an implementation perspective, two lessons emerge. First, NbS design reviews benefit from explicitly separating "what is measured" (reach-scale water-surface slope and friction) from "what is assumed" (vegetation drag closures), and requiring a traceable mapping between vegetation survey metrics and model resistance inputs. Second, adopting a staged workflow—screening with conservative resistance bounds, followed by targeted full-scale validation for high-consequence reaches—reduces approval risk while avoiding overdesign driven by uncertain vegetation parameters. Third, presenting outcomes as decision-ready options (water level and conveyance ranges across plausible vegetation states) improves communication with stakeholders and supports adaptive vegetation management plans tied to monitoring.

ABSTRACT. The global water resource crisis is becoming increasingly severe. The traditional governance model dominated by engineering technology is difficult to fundamentally solve the problem of imbalance in the relationship between humans and water. As a major country in water resource management, China has come to realize through long-term water governance practices that the essence of river and lake issues lies in the ethical relationships between people and between people and nature. The River and Lake Chief System, which was fully implemented in 2016, is not only an innovation in management mechanisms but also a profound transformation in water management concepts, marking an ethical shift from "conquering rivers" to "respecting rivers". This system, through a five-level system of provincial, municipal, county, township and village, has incorporated over 1.2 million river and lake chiefs into the governance network, achieving systematic management of 45,000 rivers and 1.51 million kilometers of river channels across the country. Its innovation lies in integrating ethical values into institutional design, providing a practical path for building a modernization path of "harmonious coexistence between humans and nature". The River and Lake Chief System is a major institutional innovation in the field of water governance for China's ecological civilization construction. This article, from the perspective of the combination of ethics and institutions, systematically analyzes how the river and lake chief system effectively addresses the "tragedy of the Commons" predicament in river governance through the shift in ethical values and institutional design innovation. The article delves deeply into the three ethical dimensions of responsibility ethics, river morality and river justice contained in the river and lake chief system, and explains the significance of its paradigm shift from technical governance to ethical governance. Based on China's governance practices, this paper demonstrates the theoretical value and practical achievements of the river and lake chief system in building a community of life for humans and rivers, providing Chinese wisdom for global water governance.

ABSTRACT. 1. Introduction As a water engineer, there is a professional and ethical responsibility to develop comprehensive solutions that consider the full context of a given problem or project, especially the interests of the communities and stakeholders it serves. In Aotearoa New Zealand, this includes acknowledging the role of mana whenua (people of the land) and their perspectives.

This study seeks to provide a contextual understanding of Te Ao Maori in the water engineer space to inform engineers about culturally aligned engineering practice in Aotearoa New Zealand.

2. Understanding Te Ao Maori The foundational concepts within Te Ao Maori that can help contextualise the perspective of Maori/mana whenua are tapu, noa and mauri, which are reinforced through whakapapa, kaitiakaitanga and tikanga maori.

3. Maori Perspectives on Water and Wastewater treatment Projects Three case studies were investigated for analysis of Maori perspectives in water treatment projects to interpolate common themes, concerns and concepts. Pukekohe Wastewater Treatment Plant upgrade, Gisborne District Council’s (GDC) Main Wastewater Consent Project, and Rotorua Wastewater Treatment Plant upgrade.

4. Current engagement between engineering & Mana whenua The Resource Management Act 1991 (RMA), specifically Part 2, Section 6(e), recognises the relationship of Māori and their culture and traditions with their ancestral lands, water, sites, wāhi tapu, and other taonga as a matter of national importance. This provision mandates consultation and acknowledges the special position of Māori as tangata whenua.

5. Conclusion To advance culturally aligned water engineering practice in Aotearoa New Zealand it is important to frame the way water is conceptualised. A move from viewing water as a resource for human utility to intrinsically recognising it as a living entity with mauri, supports approaches that preserve ecological balance.

Early and ongoing engagement with mana whenua is ideal and ensures that cultural priorities are embedded in the foundation of project planning. Relational engagement, grounded in trust and transparency, is effective and be sustained throughout the lifecycle of the project.

Land-based systems and natural treatment processes, such as constructed wetlands or rapid infiltration beds, offer preferred solutions that allow for both physical and spiritual reconciliation of wastewater. While these methods may not always offer the fastest or most cost-efficient solution, they uphold the cultural integrity of the water and reinforce the importance of returning it to the environment in a restored state.

ABSTRACT. People in flood prone area in Japan constructed traditional self-defense structures for flood inundations around a residence in 17 to early 20 Century. In case of Mizuka, the ground elevation was heightened up to levee height or river-bed level which is higher than the neighboring plain, and a warehouse was constructed on the mound which can be utilized for vertical evacuation at floods. In the Yodo River Basin, warehouses called ‘Dangula’ were constructed using high stone-wall foundation and stand for flood inundations. In the Ohi River Basin, the shape of residence to river direction was set like a ship or triangle for repelling the inundation current (Ship-shaped residence). In the Arakawa River, moats surrounding Mizuka called Kamaebori existed. The role is assumed to guard from floods like the ship-shaped residence; however, many unknown points existed. Especially, large and different types of Kamaebori existed in the riverside and at river confluences. For clarifying the role, the shape of the Kamaebori in the Arakawa River Basin was classified ‘Front type (FT)’, ‘L-shape type (LT)’ and ‘Horseshoe type (HT)’ according to one, two and three directions of the residence, respectively. In addition, the flood inundation characteristics, especially around the locations where Kamaebori existed, were investigated using two-dimensional depth-averaged flow model using old river course data. 2-years return period of flood was selected for simulation as the levee height was low (1-3 m) at that time, and flow capacity was low. Numerical simulation demonstrated that Kamaebori faces the complex directions corresponding to the flood inundation currents. The LT and HT correspond to two to three directions of flood approaching direction affected by backwaters and land elevation differences affected by natural levee. Difference between Mizuka with Kamaebori (MK) and without Kamaebori (MN) also elucidated. MK existed in high flood inundation velocity and fluid force locations compared with MN. It shows that Kamaebori has a role to repel the current like a ship-shaped house. As Mizuka itself has a role to survive from stored type floods, it can be concluded that MK is a hybrid type self-defense structure for residence.

ABSTRACT. This study analyzes the historical evolution of the human-river relationship in Korea and proposes a framework for establishing River Ethics as a core mechanism for future water sustainability. The central issue addressed is the persistent socio-ecological conflict and vulnerability resulting from the industrial development model. Korea provides a unique case study of rapid cultural transition, evolving from Phase 1: Adaptation (Ancient→Joseon), characterized by implicit ethics of reverence and reliance on natural rivers for navigation, to Phase 2: Control (1900s→1980s). This Control phase, driven by the hydraulic mission for water utilization and flood control, prioritized economic growth, leading to ecological degradation. The subsequent Phase 3: Restoration (1990s→Present), symbolized by Cheonggye stream restoration, successfully re-established public amenity but often failed to fully address the river’s intrinsic ecological rights. This work employs Groenfeldt's five-value framework of water ethics (Environmental, Social, Economic, Cultural, Governance) to interpret these transitions. The analysis reveals that engineering, legal, and institutional solutions lacking ethical foundations cannot resolve deep-seated conflicts. The study concludes that Korea must transition into Phase 4: Ethical Coexistence. This requires adopting explicit River Ethics by integrating five core values: 1) Environmental Value (acknowledging the river's inherent rights, e.g., estuary restoration); 2) Social Value (ensuring water justice through equitable burden/benefit sharing); 3) Economic Value (shifting focus from supply-oriented growth to efficiency and conservation); 4) Cultural Value (prioritizing the spiritual and mental connection with water over physical amenity); and 5) Governance Value (embedding procedural ethics via multi-stakeholder participation). Establishing this new culture aligns traditional wisdom with modern water policy, offering a sustainable transition model.

ABSTRACT. In recent years, in the context of promoting river basin flood control, the utilisation of farmland as flood storage areas has gained attention. However, it is difficult for most of farmers to accept flood storage area plans. Consequently, challenges in reaching consensus have been pointed out. It is known that prior to the implementation of modern flood control works, such as continuous embankment improvements, flood control was carried out on the premise that water would overflow, aiming to minimise the damage. In those days, agriculture was practised on the assumption that floods would occur. There ware a risk of flood, but coincidently, it is known that there were also the benefits of flooding, such as the influx of fertile soil from upstream. However, with the spread of modern flood control, agriculture itself transformed, becoming systematised on the premise that flooding would not occur. There is a discrepancy between river administration, which promotes basin-wide flood control, and agricultural administration, which promotes industrialised agriculture. Therefore, it is hoped that understanding the agriculture that coexisted with floods, once seen in Japan, can reduce the barriers to introducing watershed flood control. This study therefore focuses on agricultural specialities cultivated in areas frequently affected by flooding, aiming to identify the symbiotic relationship between agriculture and floods in various locations. To achieve the objective, speciality products were extracted from literature compiling nationwide specialities published between the 1950s and 1980s, under the following conditions: 1) Produced utilising the region's geographical characteristics, 2) Descriptions of relationships with flooding, rivers, or soil, 3) Excluding rice, mulberry, or tobacco, which were nationally encouraged crops. Of the 43 items extracted, only three explicitly mentioned a connection to flooding. Several additional cases with clear links to flooding were included in the analysis. Descriptions of symbiotic relationships were gathered from relevant literature, local histories, and research papers for each case, whilst the topography and soil conditions of cultivation sites were also examined. In conclusion, it is clear that from a short-term perspective, flooding constitutes a risk factor necessitating flood countermeasures and the selection of appropriate crops and varieties. Conversely, from a long-term perspective, it is also evident that the soil formed by rivers and floods can confer benefits due to its characteristics and the thickness of its sedimentary layers. Agriculture in regions frequently affected by flood damage can be said to be sustained based on this balance between risk and benefit.

ABSTRACT. International exchange programs can accelerate the formation of next‑generation water professionals by combining shared problem framing, cross‑cultural teamwork, and practice‑oriented numerical modeling. This presentation introduces HydroAsia, an academic exchange program that fosters collaborative learning among leading water‑related universities and institutes worldwide, aiming to explore solutions to local and global water challenges while sharing effective water‑management technologies and knowledge. First hosted by Incheon National University in 2009, HydroAsia reached its 17th edition in 2024 and has become a key regional platform for joint education and cooperation among professors and undergraduate/graduate students. In 2024, approximately 80 participants from 14 countries joined the program, involving multiple universities and research institutes across Asia and beyond. HydroAsia’s core theme is flood mitigation, supported by numerical simulation tools such as SWMM, MIKE, and the HEC series, with an additional theme of ecological river restoration. The program blends online preparatory activities with an intensive on‑site module that includes field trips, team‑based projects, numerical model construction, scenario simulations, and solution‑oriented presentations. To illustrate the training workflow, we present two Incheon (Republic of Korea) case modules used in the program. The learning sequence integrates (a) local context reading, (b) ethics‑guided problem framing, and (c) model‑based scenario testing. For Seunggicheon (Seung‑gi Stream), students examine an urban watershed (31.96 km²; stream length 10.47 km) where maintenance water is supplied from a wastewater treatment plant (35,000 tons/day) and water‑quality improvement actions include ecological stream construction and sludge removal. These observations are translated into hydraulic questions (e.g., flood water levels and mitigation options) and explored through river simulation exercises. For Guwol‑dong, a dense downtown district with a high floating population, students conduct urban flood analyses using SWMM. A sewer network model (158 pipes) is tested using observed rainfall from the 23 July 2017 flood event and extended to frequency‑based design storms (e.g., 50–200‑year conditions) to compare mitigation strategies. Based on these modules, we discuss future directions to strengthen HydroAsia’s contribution to educating young hydraulicians: (i) standardizing reproducible modeling packages (curated datasets, version‑controlled model files, and reporting templates); (ii) strengthening traceability between field evidence, modeling assumptions, and decision accountability to improve communication of uncertainty and trade‑offs; and (iii) adopting competency‑based assessment (pre/post diagnostics and rubric‑based evaluation) to document learning gains. We conclude that a field‑based exchange “living lab” can provide a scalable pathway for developing globally minded, practice‑ready young water‑resources engineers in Asia.

ABSTRACT. Comprehensive evaluation of ecosystem health (EH) is crucial for estuarine and deltaic wetlands facing intensifying natural and anthropogenic stresses. However, the impact of geomorphological evolution, a key inherent process that drives the evolution of estuarine and deltaic wetlands has not yet been systematically addressed. This review synthesizes 216 publications to elucidate how geomorphological processes affect three EH dimensions: vigor, organization, and resilience. Bibliometric analysis reveals a rapid growth in the research endeavors since 2014, with a geographic focus on estuarine and deltaic wetlands in the USA, China, the Netherlands, Germany and Italy. The research focus has evolved from geomorphological processes to system resilience under external disturbances, especially sea level rise. Supported by the large language model, our analysis further reviewed the effects of depositional pattern and geomorphic unit formation on all three dimensions of EH, highlighting contrasting effects due to scale-dependent effects and species-specific environmental preference. Moving forward, we recommend that future research needs to focus on three directions to advance the comprehensive assessment and effective restoration of EH in estuarine and deltaic wetlands: 1) enhancing observation via multi-source data fusion, 2) improving the understanding of the bio-geomorphological interactions and advanced coupled models, and 3) optimizing EH assessments through an adjusted Pressure-State-Response framework.

ABSTRACT. South Korean rivers have undergone a substantial transformation from "white rivers," dominated by exposed Gravel and/or sand bars, to "green rivers" with extensive vegetation cover. Approximately 60% of rivers have experienced this transition since the 1980s. This white-to-green river transition poses significant challenges to flood control capacity and aquatic habitat conservation. However, the causal mechanisms remain poorly understood due to limited spatial coverage and a lack of comprehensive analysis of multiple factors. This study quantified vegetation expansion across 17 major national rivers (total of 85 reaches, ca. 1-3 km each) using aerial photographs from the 1970s to the 2020s and analyzed four potential factors: flow regulation by dams, dredging, nutrient loading, and climate change. Rivers were categorized into regulated (n=9, with upstream dams) and non-regulated (n=8, without dams) systems. For nutrient loading, we analyzed institutional changes in livestock farming policies and temporal patterns of total nitrogen and total phosphorus from national water quality monitoring data. Climate change effects were evaluated by examining altered precipitation patterns, particularly spring rainfall increases and summer rainfall decreases. Dredging activities were identified through aerial photograph interpretation. Both regulated and non-regulated rivers exhibited substantial vegetation expansion after 2010, with vegetation area increasing three- to tenfold compared to the 1970s. Notably, although major dams in Korea were constructed primarily between the 1960s and 1980s, this widespread vegetation expansion emerged decades later in the 2010s. This temporal discrepancy indicates that dam construction alone cannot fully explain the recent transition and that multiple interacting factors contributed to the overall change across different river types. This study reframes the white-to-green river transition from a single-factor, dam-centric explanation to a multi-factor framework. Using a structural equation modeling approach, we will quantify the direct and indirect effects of each factor on vegetation expansion. By identifying how multiple factors—including altered flow regimes, nutrient enrichment, climate shifts, and management interventions—interact to drive vegetation expansion, this framework establishes a comprehensive understanding of the transition mechanisms. These findings provide an empirical basis for developing adaptive river management strategies that integrate flood control, water quality, and ecological conservation at the basin scale.

ABSTRACT. Approximately 40% of aquatic plants in Japan are on the Red List of the Ministry of the Environment, and their conservation is considered an urgent priority. Because rivers are the second largest freshwater bodies in Japan after lakes and marshes, major rivers are also considered important habitats for aquatic plants, but there is less information on them than on lakes and marshes. Therefore, the purpose of this study is to characterize the aquatic plants growing in Japan's major rivers using data from the National Census on River Environments (NCRE). And comparing 2 cycles (FY2001-2005: 3-cycle and FY2006-2015: 4-cycle), how did aquatic plants’ occurrence change in major rivers? Some survey sites didn’t overlap. Therefore, we use only overlapping sites for analysis. The relevant aquatic plants were extracted from the NCRE data, and information on the characteristics of each species (total number of life forms, native or alien species, and life history information) was assigned. As a result, 271 rivers (including 104 river basins) were used for analysis from the 2-cycle NCRE surveys. The number of plant species was high, the number of survey sites was high, and the length of these rivers is long. Aquatic 2 species appeared at a high number of sites in both the 3 and 4 cycles. Many species in Aquatic 2 use Japan's major rivers as their habitat. Extract only disappeared species in 4 Cycle, Percentage of aquatic 1’s extinction was higher than other categories. And 7 species were listed in the Red List in Japan. And aquatic 2, no species were extinct. Although the occurrence was only in 4 cycles, many native species were included. Many species, including Aquatic 2 (Emergent species), use Japan's major rivers as their habitat. Most of the species that disappeared in the fourth round were Aquatic 1, but no species of Aquatic 2 disappeared. Aquatic 2 was thought to be growing stably. On the other hand, there were also many species that only appeared in the 4 Cycles. Japan's major rivers are stable habitats for emergent aquatic plants and are also important freshwater environments where many endangered species grow.

ABSTRACT. Vegetation-induced modifications of flow and sediment transport play a critical role in channel morphodynamics, yet sediment sorting processes in vegetated channels with mixed-size sediment remain insufficiently understood. This study experimentally investigates the effects of riparian vegetation density on channel geomorphic evolution and surface sediment sorting in an alluvial channel composed of mixed sediment. Laboratory experiments were conducted in a 12 m-long adjustable flume, where natural vegetation (alfalfa) was cultivated at different stem densities. Temporal variations in channel morphology, sediment discharge, and surface grain-size distribution were systematically measured and analyzed. The results indicate that increasing vegetation density significantly alters flow structure by reducing velocity within vegetated zones and inducing boundary-layer flows between vegetated areas and the main channel. These flow modifications promote sediment trapping within vegetation zones and reduce sediment discharge at the downstream end. Surface sediment sorting exhibited pronounced spatial variability under vegetated conditions. With increasing vegetation density, the dimensionless median grain size of the surface layer decreased, while the relationship between bed elevation change and grain-size variation became increasingly irregular. In highly vegetated cases, sediment fining occurred despite local bed degradation, reflecting the dominant influence of vegetation-induced flow redirection and sediment capture. Statistical analysis of time-series bed elevation data further revealed that moderate vegetation density stabilized low-flow channels, whereas higher vegetation density enhanced channel instability through intensified channel migration and reorganization. These findings demonstrate that riparian vegetation density exerts a non-linear control on sediment transport efficiency, surface sediment sorting, and channel stability in mixed-sediment rivers. The results provide experimental insights into vegetation–sediment–flow interactions that are essential for predicting morphodynamic responses in vegetated river systems and for informing river restoration and management strategies.

ABSTRACT. The transition from engineered "white" rivers to ecologically functional "green" rivers represents a fundamental shift in river management philosophy, particularly under the context of climate change adaptation. Central to this transition is understanding how vegetation patches modify river morphology and create resilient channel-floodplain systems. Concurrently, the field has witnessed a paradigm shift in measuring fluvial morphodynamics—from labor-intensive point-based surveys to high-resolution three-dimensional techniques. Among these, Terrestrial Laser Scanning (TLS) has emerged as a powerful tool, enabling millimeter-scale topographic measurements over extensive areas and capturing detailed bed evolution that was previously unattainable. This study investigates morphological changes induced by vegetation patches through large-scale flume experiments utilizing TLS-based high-resolution topographic monitoring. Vegetation patch configurations of varying density and spatial arrangement were installed in a mobile-bed flume, and bed topography evolution was captured using repeated TLS surveys at multiple time intervals, generating precise digital elevation models for quantitative change detection. Results demonstrate that vegetation patches significantly alter local flow structure, creating distinct zones of erosion and deposition. Flow acceleration around patch margins promotes scour, while reduced velocities within and downstream of patches facilitate sediment accumulation. The magnitude of bed elevation change correlated strongly with patch density and the ratio of patch width to channel width. Staggered patch arrangements induced more complex but stable bed morphology compared to aligned configurations. The biogeomorphic feedback mechanism—whereby vegetation establishment modifies topography, which subsequently influences further colonization—was clearly observed, suggesting a self-reinforcing process critical for long-term channel evolution. These feedback processes are essential for developing self-sustaining green river systems capable of adapting to altered flow regimes driven by climate change. These findings provide quantitative relationships for predicting morphological trajectories following vegetation establishment, supporting climate-adaptive river management strategies that embrace nature-based solutions.

Acknowledgement: Research for this paper was carried out under the KICT Research Program (project no. 20250223-001, Research on Establishing a Foundation for Responding to Current Issues and Challenges in Water Management) funded by the Ministry of Science and ICT.

ABSTRACT. Flow resistance in partly vegetated channels is governed not only by vegetation density but also by the spatial arrangement and vertical structure of vegetation patches. Previous studies suggest that, particularly under low-velocity conditions, patches located near channel banks can yield lower reach-scale flow resistance than comparable patches placed along the centerline. In addition, for woody vegetation patches with similar blockage, the presence of herbaceous understory beneath the canopy can alter flow resistance and associated flow structures. Therefore, this study quantifies how lateral patch placement (centerline versus bankside) and within-patch structural configuration (woody-only, herbaceous-only, and combined woody–herbaceous) affect water-surface slope, bulk flow resistance, streamwise velocity structure, and turbulent kinetic energy (TKE) in a partly vegetated channel, using large-scale experimental measurements. Experiments were conducted in an outdoor trapezoidal channel (bottom width 3 m, top width 11 m, and bank slope 1:2), focusing on a ~70 m instrumented reach with a bed slope of 1/800 and a movable bed with a median grain size of 1.06 mm. Discharge was controlled between 1.4 and 2.5 m³/s, and flow depth was maintained at approximately 0.70–1.10 m. The patches with the same number of plants and blockage ratio were laterally relocated between the centerline and bankside positions to vary patch placement. To assess vertical/within-patch structural effects, patches with identical planform area were arranged as woody-only, herbaceous-only, or combined woody–herbaceous conditions. Water-surface slope was measured using pressure sensors, discharge was measured with ADCP, and channel geometry was surveyed by terrestrial laser scanning. Streamwise velocity and turbulent kinetic energy were computed from three-dimensional velocity components measured with an ADV at mid-depth (0.5 of flow depth). Compared with the centerline patch configuration, the bankside patches reduced the bulk flow resistance coefficient (friction factor, f) by approximately 16%, with the most significant reduction, about 25%, observed under low-discharge, low-velocity conditions. The centerline-patch configuration exhibited greater lateral TKE variability, quantified as the max–min difference, than the bankside-patch configuration, indicating stronger shear production and more complex mixing. Vegetation structural configurations with understory growth exhibited steeper friction slopes than those without understory and showed approximately 31% higher bulk friction factors across the experimental conditions, with a slightly larger relative increase (within ~2%) at higher discharges.

Acknowledgement: This research was funded by the Korea Environment Industry & Technology Institute (KEITI) through the Climate Resilient R&D Project for Water-Related Disaster Management, funded by the Korea Ministry of Climate, Energy, Environment (MCEE)(RS-2022-KE002091).

ABSTRACT. Abstract The dew point is a representative physical indicator of atmospheric water vapor content and is thermodynamically closely linked to air temperature. Recent sustained increases in temperature driven by climate change are inducing fluctuations in associated dew points, increasing the likelihood of trends diverging from historical patterns. In particular, the maximum dew point is a key variable determining the scale of extreme rainfall events, directly influencing the estimation of Probable Maximum Precipitation (PMP). Since PMP serves as the design criterion for large-scale hydraulic structures such as dams, ensuring the safety of domestic infrastructure requires a precise analysis of long-term variations in the maximum dew point and the identification of abrupt change points. Therefore, this study utilizes the Bayesian Estimator of Abrupt change, Seasonality, and Trend (BEAST) algorithm to analyze the variability characteristics of domestic maximum dew point time series. Prior to the algorithmic application, efforts were made to minimize numerical distortions arising from differences in the temporal resolution of observational data during the construction of the 12-hour maximum dew point time series. To this end, correction factors based on observational temporal resolutions were introduced to enhance the physical consistency and analytical precision of the data.

Acknowledgments This work was supported by Korea Environment Industry & Technology Institute(KEITI) through Research and Development on the Technology for Securing the Water Resources Stability in Response to Future Change Project, funded by Korea Ministry of Climate, Energy, Environment(MCEE)(RS-2024-00332300).

ABSTRACT. Recent climate change has intensified abnormal weather events, including more frequent heatwaves and greater precipitation variability, raising concerns about both drought frequency and the severity of associated damages. Unlike floods, droughts are difficult to define in terms of onset and termination, and assessments of drought intensity and duration vary across observational datasets and climate model outputs. Consequently, drought-disaster vulnerability assessments based on a single index may bias interregional prioritization. Although North and South Korea share broadly similar natural environments, drought impacts can differ substantially due to contrasting socio-economic systems and management conditions. This study therefore evaluates and compares drought vulnerability across North and South Korea using multiple drought indices under future climate change scenarios. Global climate model outputs from CMIP6, based on the Shared Socioeconomic Pathways (SSPs) presented in the IPCC Sixth Assessment Report (AR6), are employed, with the high-emissions scenario SSP5-8.5 used to characterize extreme future conditions. To represent key hydroclimatic drivers—including precipitation, temperature, and atmospheric evaporative demand—we apply the Standardized Precipitation Index (SPI), Standardized Precipitation Evapotranspiration Index (SPEI), self-calibrating Effective Drought Index (scEDI), and Evaporative Demand Drought Index (EDDI). In addition, the Reconnaissance Drought Index (RDI), which captures the relative imbalance between precipitation and potential evapotranspiration, is included to broaden the interpretation of drought characteristics. By comparing and synthesizing index-specific results, spatial patterns of drought vulnerability are quantified and contrasted between North and South Korea, and priority regions for future agricultural drought response are identified. The proposed multi-index framework reduces sensitivity to any single indicator and provides a robust, scenario-consistent basis for regional prioritization, drought-risk management, and climate adaptation planning. This work was supported by Korea Institute of Planning and Evaluation for Technology in Food, Agriculture, and Forestry (IPET) through Intelligent Agricultural Infra Management for Climate Change Development Program, funded by Ministry of Agriculture, Food and Rural Affairs (MAFRA) (RS-2025-02219764)

ABSTRACT. This study developed a data-driven framework to project future domestic water demand across 116 municipalities in South Korea under evolving socio-economic and climatic conditions. Existing national water-management planning methods rely heavily on aggregated indicators and static assumptions, resulting in substantial discrepancies between estimated and actual domestic water use. To address these limitations, this study reconstructed unit water-use rates by integrating historical consumption records with demographic and supply-coverage data, and enhanced future projections through a Long Short-Term Memory (LSTM) deep-learning model optimized via Bayesian methods. The optimized LSTM model demonstrated strong predictive performance, with validation MASE values below 1.0 for most municipalities, indicating improved accuracy over baseline time-series models. Future unit-use trajectories differed substantially across regions: metropolitan areas such as Seoul and Daejeon showed stabilization after historical declines, whereas emerging cities like Sejong displayed sustained increases linked to rapid population growth. These reconstructed unit-use rates, combined with SSP-based demographic projections and water-supply coverage ratios, enabled the estimation of annual municipal domestic water demand from 2022 to 2100. Scenario-based projections revealed pronounced spatial and temporal divergence. For example, Incheon was projected to experience a 17–29% decline in annual domestic demand depending on the scenario, driven primarily by depopulation despite modest increases in per-capita water use. In contrast, Sejong showed demand increases of 22–40% due to substantial population growth. The framework also converted annual demand into monthly and sub-district scales using seasonal water-use factors and 2020 population shares, enabling finer-resolution drought-vulnerability and resource-allocation analyses.

ABSTRACT. Background: Hydrometeorological disasters frequently strike the northern zone of Bangladesh, causing devastating impacts on the lives and livelihoods of communities. Purpose: The objective of this study was to investigate how hydrometeorological disasters affect the lives and livelihoods of people living in the riverine areas of northern Bangladesh. This study was conducted in flood- and riverbank-erosion-prone areas of Kurigram district, Bangladesh. Methods: The qualitative research employed a phenomenological approach. Primary data was collected using in-depth interviews and focus group discussions (FGDs) with local communities and key informant interviews with local leaders, agricultural experts, and fisheries experts. The results are interpreted using sustainable livelihoods frameworks and build-back-better frameworks. Results: The study revealed that hydro-meteorological disasters, such as floods and riverbank erosion during monsoon season, cold spells in winter, and seasonal droughts, pose major problems to local communities in terms of their everyday lives, affecting their livelihoods. Four themes emerged from the analysis of in-depth interviews and FGDs with local communities, and they are, firstly, the impacts of hydro-meteorological disasters on the socio-economic conditions of people; secondly, the impact on agricultural production; thirdly, the impact on health status; and finally, the impact on water availability. The findings of this study suggested that hydrometeorological disasters have serious consequences for the livelihoods of riverine communities, ultimately hindering their ability to rebuild effectively. Conclusion: The findings of this investigation provide clarity regarding the priorities associated with strategies for reducing flood risk in riverine areas. A risk-based approach to actual livelihoods is necessary to ensure the sustainability of riverine communities.

ABSTRACT. Compound heatwave and extreme precipitation (CHWEP) events pose much greater risks to communities compared to individual climate extremes. While recent studies have focused only on increasing CHWEP events, revealing underlying physical mechanism under climate change, their long-term and seasonal characteristics in South Korea have not quantitatively analyzed. This study investigates the frequency and intensity of CHWEP across South Korea over the past four decades (1986–2025), with particular emphasis on evolving seasonal patterns in recent decades (after 2000). In addition, we contextualize the record-breaking CHWEP events of 2024–2025 within this 40-year historical period. We show that CHWEP events in 2024–2025 far exceeded historical variability in terms of both intensity and frequency, signaling an emerging climate transition. We also identify statistically significant increases in the CHWEP frequency during fall and winter, suggesting a “seasonal expansion” of such compound hazards beyond its historical primary season (i.e. summer) due to monsoon climate. These findings highlight a growing risk of CHWEP events outside the traditional hazard season and underscore the need for adaptive strategies to manage these evolving multi-hazard climate impacts in South Korea

This work was supported by National Research Foundation (NRF) grants of Korea funded by the Korean government (MSIT) (RS-2022-NR072388). This work was supported by Korea Environment Industry & Technology Institute (KEITI) through Citizen driven climate change adaptation policy adaptation by policy evaluation using citizen-socio-economic-environmental data (2025-071(R))Project, funded by Korea Ministry of Climate, Energy and Environment (MCEE) (grant number: RS-2025-02263293)

ABSTRACT. To ensure the stability of agricultural water supply, the Korea Rural Community Corporation (KRC) constructed five underground dams under the 10-Year Agricultural Water Resources Development Plan (1982–1991), which are currently in operation. In response to persistent drought conditions, this study aims to analyze the drought characteristics of an underground dam watershed by developing a hydrological model for the Okseong Underground Dam watershed, operated by KRC, to support integrated surface water–groundwater monitoring using groundwater resource development facilities. The study area was defined as the Okseong Underground Dam watershed, located in the downstream section of the Yugu Stream within the Yugu Stream basin in Gongju-si, Chungcheongnam-do, Republic of Korea. For hydrological analysis of the underground dam watershed, the Soil and Water Assessment Tool (SWAT) model was selected, as it is capable of simulating surface runoff, interflow, and baseflow processes. Model input data, including soil maps, land use/land cover maps, digital elevation models (DEM), hydrological data, and meteorological data, were collected and processed. Model calibration and validation were conducted using observed streamflow data from the Gukjae Bridge gauging station on the Yugu Stream, incorporating wet, normal, and dry years. Based on the analysis of groundwater discharge behavior in the underground dam watershed, this study evaluates the drought characteristics of the watershed.

This work was supported by Korea Environment Industry & Technology Institute(KEITI) through Climate Resilient R&D Project for Water-Related Disaster Management, funded by Korea Ministry of Climate, Energy and Environment(MCEE).(2022003610002)

ABSTRACT. Climate change is expected to intensify hydroclimatic and coastal hazards, increasing the risk from extreme ocean waves and challenging the stationary assumptions traditionally used in coastal and port design. However, reliable estimation of future extremes remains difficult because wave projections inherit systematic errors from both the forcing Global Climate Models (GCMs) and physical numerical limitations of wave models. While bias correction is widely applied to reduce these discrepancies, conventional empirical quantile mapping (EQM) often shows limited skill for the upper tail, and may underestimate extremes above the 99th percentile. This study proposes an extreme-oriented bias correction framework for long-term wave projections in the Northwest Pacific (20°–43°N, 117°–142°E), including the Korean coastal region, that combines the strengths of EQM with a hybrid bias correction framework specifically designed to place stronger emphasis on extreme values. Wave parameters are simulated using the third generation spectral wave model SWAN forced by six horizontally high resolution HighResMIP GCM wind fields under the SSP5-8.5 scenario. We consider a historical period (1985–2014) for calibration and evaluation and a future period (2015–2050) for projection. The proposed hybrid transformation applies an empirical Gumbel quantile mapping (EGQM) based on a Gumbel distribution for values above the 99th percentile, and subsequently performs an independent EQM-based correction on exceedances beyond the 99th percentile, thereby reducing systematic underestimation of extreme waves while avoiding strong parametric tail assumptions. Following a cross-validation-based calibration strategy, we assess performance using quantile based skill metrics and design relevant extreme indices. The resulting bias corrected projections provide more consistent estimates of extreme waves required for coastal and port engineering applications and support non-stationary assessments of future design conditions under a high emission climate scenario.

ABSTRACT. The pressing risk and management posed by cyanobacterial blooms in worldwide lakes is inevitably influenced by structure characteristics of cyanobacterial community, the uncertainty of which might be extensively heightened by frequent and extreme regional climate conditions in recent decades. Previous research has increased our understanding of cyanobacterial community dynamics and succession at the single-lake or local scales, however, the distribution characteristics and the dominant influence factors of cyanobacterial community structures and ecological strategies in lakes at the global or large-spatial scales remain unknown. To address this gap, we employed the multistable state theory and survival strategy sorting to explore the community structure states of cyanobacterial blooms and the dominant factors in north temperate lakes via a collected dataset containing 1665 samples from 1490 lakes in 2001-2022, and then predicted the variation of cyanobacterial community structure states of 966,411 lakes in north temperate zone during 2020-2100 under RCP60 and RCP85 scenarios separately. The results showed that scum genera of cyanobacteria (top-2: Microcystis and Anabaena) accounted for 51.5% of the biomass across 1490 lakes, followed by dispersive genera (43.6%, represented by Planktothrix and Aphanizomenon). In contrast, dispersive genera (top-2: Chroococcus and Aphanocapsa) accounted for 68.4% of the occurrence frequency, followed by scum genera (23.3%, top-2: Anabaena and Microcystis). Based on canonical correspondence analysis of community structures at the genus level (indicated via CCA1), the collected samples displayed significant bistable states (A and B, with CCA1 = 0.25 as the cutoff) dominated by dispersive (54.7%) and scum (42.7%) genera, respectively. As the most pronounced factor, discharge regulated by local meteorological conditions displayed a significantly influence on CCA1, with higher discharge values favoring state A. By an optimal random forest algorithm, the predicted CCA1 value of 966411 lakes in summer exhibited a U-shaped pattern from 2020 to 2100 under two scenarios, with the rock-bottoms in 2050 (RCP60) and 2070 (RCP85) separately. Under both rock-bottom conditions, the top-abundance state A and B are mainly distributed in 50° ~ 65°N and 55° ~ 65°N, respectively. Compared to 2020, though the hot zone of bidirectional shifts between two states consistently occurs in 50° ~ 65°N, the high-intensity zone of regime shift from state B to A will move from 60°N under RCP60 to 25° ~ 45°N under RCP85. The findings in this study put novel insights into the importance of climate and hydrology change on cyanobacterial community structure states in north temperate lakes.

ABSTRACT. Conventional drought assessments focus on how often drought occurs, but fail to capture how drought evolves once initiated. This study examines the temporal dynamics of soil moisture droughts across the Godavari River Basin, India, using a grid-wise Markov transition probability framework applied to 42 years (1981 to 2022) of satellite-derived soil moisture from the European Space Agency Climate Change Initiative Soil Moisture product, version 0.81 (ESA CCI SM v0.81). Monthly soil moisture time series were used to compute the Standardized Soil Moisture Index (SSI) at each grid cell, and drought events were identified through run theory (SSI below -1). Events were classified into four discrete severity states, and first-order Markov transition matrices were constructed to quantify drought persistence, escalation, and recovery probabilities. Results show that basin-mean persistence, escalation, and recovery probabilities are 0.25, 0.26, and 0.15, respectively. Escalation exceeds recovery over 85% of the basin area. Persistence-dominated regimes cover 56.7% of the basin and escalation-prone areas cover 36.1%, with recovery-dominated conditions confined to 7.3% of the basin. A composite Soil Moisture Drought Memory Index (SMDMI) yields a basin-mean value of 0.78, with a standard deviation of 0.10, confirming moderate-to-high drought memory across most of the basin. These results provide a transition-based framework for quantifying drought propagation and improving drought risk assessment in water-stressed river basins.

ABSTRACT. Indonesia’s mangroves face repeated hydroclimatic shocks linked to ENSO variability and extreme rainfall and flooding. At the same time, many river basins have been increasingly regulated through dams and check dams. These structures can alter downstream freshwater and sediment delivery that may influence mangrove recovery. This study compares post-event mangrove recovery across selected Indonesian river–coast systems by using an event-aligned analysis. We select several comparable watersheds with similar size and coastal settings but differing regulation intensity, including large reservoirs and check dams. Mangrove canopy condition is monitored from 2017 to 2025 using Sentinel-2 observations within a fixed mangrove domain. Canopy condition is represented by a compact set of optical indices capturing complementary stress pathways: vegetation vigor (NDVI, EVI), chlorophyll-sensitive response (NDRE), and moisture-related stress (NDMI). These indicators are combined into a composite condition score and expressed as standardized anomalies relative to each location’s normal seasonal cycle. Hydroclimatic shock periods are identified using ENSO phase windows and independently detected rainfall and flood anomalies from gridded climate products. To ensure comparability across regions, events are defined and grouped by similar canopy-impact magnitude, and recovery trajectories are aligned by the timing of maximum impact. For each event and watershed-associated coastal segment, we estimate impact and recovery metrics, including maximum condition deficit, stress duration, time-to-recovery, and recovery rate from post-impact rebound curves. The novelty is a consistent multi-watershed design that links event-scale mangrove recovery trajectories to contrasting river regulation settings. Expected outputs include comparable recovery-rate profiles across watershed groups and identification of river–coast systems where recovery is systematically slower or faster under similar shock intensity. These findings provide evidence to guide subsequent inland–coastal coupling analyses and targeted field validation of the role of river regulation in shaping coastal ecosystem resilience.

Acknowledgement This work was supported by the Korea Environmental Industry and Technology Institute (KEITI) through the Wetland Ecosystem Conservation Technology Development Project (Project No. RS-2022-KE002030), funded by the Ministry of Climate, Energy, and Environment. The first author also gratefully acknowledges the scholarship support from The Hyundai Motor Chung Mong-Koo Foundation.

ABSTRACT. Food security worldwide is increasingly becoming one of the greatest global challenges. The Tacna region, in southwestern Peru and located on the northern edge of the hyper-arid core of the Atacama Desert, is characterized by its desert and Andean zones, marking significant climatic variability. In order to maintain and improve agricultural production, especially olive cultivation, it is important to establish the crop's water stress parameters. This is the main crop in the region, and its impact on the regional economy is decisive. Important relationships such as ground surface temperature (TST) and vegetation indices (VI) were studied; these have demonstrated effectiveness in monitoring water stress in this crop, based on remote sensing information (drones and satellites) and the use of artificial intelligence algorithms. In this study, we obtained important results in our quest to identify water stress in olive groves over time, with 39% drought for the relationships (TVDINDVI and TVDISAVI), 24% severe drought (TVDINDVI), and 25% (TVDISAVI) of the studied area, compared to TVDIEVI2, which showed 37% drought and 16% severe drought. It is concluded that TVDINDVI and TVDISAVI yield better results based on the spatial definition of water stress in olive crops. They also report a wide range of options to address problems related to water stress and drought. These results are undoubtedly fundamental for future water resource management in the olive sector in arid areas of southern Peru, especially the Tacna region.

ABSTRACT. Soil moisture is commonly used to represent catchment wetness conditions and to interpret hydrological responses under drought and climate variability. However, runoff responses often exhibit substantial variability even under similar soil moisture conditions, suggesting that soil moisture alone may not fully capture the mechanisms controlling runoff generation. This study assesses how runoff responses under comparable soil moisture conditions are influenced by climate change using Shared Socioeconomic Pathway (SSP) scenarios, with a focus on process-level hydrological behavior. The Dynamic Water Resources Assessment Tool (DWAT), a process-based hydrological model, is applied to multiple catchments with contrasting hydrological characteristics. Daily DWAT outputs, including soil moisture, precipitation, total runoff, baseflow, recharge, infiltration, and actual evapotranspiration, are analyzed to examine runoff generation processes. Runoff response is quantified using event-based runoff ratios derived from simulated precipitation and discharge, enabling comparison of runoff behavior under similar soil moisture conditions across different climate settings. To ensure consistency across periods, the analysis emphasizes comparable soil moisture conditions rather than absolute soil moisture values. To evaluate climate change impacts, selected SSP-based climate scenarios are incorporated by modifying precipitation and atmospheric demand inputs. Runoff responses under historical climate conditions are compared with those under future SSP scenarios while maintaining similar soil moisture conditions. This framework allows assessment of whether climate change alters runoff responses independently of soil moisture changes and identifies the hydrological processes contributing to such differences. Results indicate that runoff responses under similar soil moisture conditions can differ markedly due to variations in subsurface flow contribution, infiltration capacity, and recharge processes. Under future SSP scenarios, the variability of runoff responses increases, reflecting enhanced sensitivity to internal process partitioning rather than soil moisture alone, particularly during dry periods.

This work was supported by Korea Environment Industry & Technology Institute(KEITI) through Water Management Program for Drought Project, funded by Korea Ministry of Climate, Energy and Environment(MCEE).(2022003610002)

ABSTRACT. This study investigates the predictability of Extreme Precipitation Days (EPDs) over the Korean Peninsula from 1981-2020 by integrating high resolution observations with large scale atmospheric and oceanic precursor signals. EPDs were defined using the 90th percentile of daily rainfall from CHIRPS and validated with APHRODITE. The spatial and seasonal climatology shows that EPDs and Very Extreme Precipitation Days (VEPDs ≥70 mm) peak sharply during July to August, particularly over central and northern regions, consistent with the mature East Asian summer monsoon. Lead-lag correlation analysis revealed that interannual EPD variability is strongly controlled by large scale processes, including enhanced precipitable water transport from the western North Pacific, reduced regional sea level pressure, strengthened low level southwestern winds, warm sea surface temperature anomalies, and elevated continental heating across East Asia. These thermodynamic and dynamic precursors provided the basis for selecting three physically coherent predictors which included: SLP (PD1), SST (PD2), and T2M (PD3). A stepwise regression model using these predictors demonstrated significant forecasting skill for July-August EPDs, achieving meaningful performance in both cross-validated reforecasts (1981-2010) and independent forecasts (2011-2020). The results highlight that seasonal EPD prediction over the Korean Peninsula is feasible when key moisture transport and circulation anomalies are incorporated. This study underscores the value of physics-based empirical modeling for improving early season forecasts of extreme precipitation, with implications for disaster preparedness and climate-risk management in South Korea under continued global warming.

ABSTRACT. Climate extremes are increasing in frequency and severity, and the co-occurrence of droughts and heatwaves creates compound events with amplified environmental and socio-economic impacts. While droughts and heatwaves have been widely studied as individual extremes, their joint behavior and spatial variability remain poorly understood, particularly at localized scales. When these occurrences coincide, they worsen water scarcity and create dangerous heat stress conditions. The frequency, severity, and dependence structure of compound drought-heatwave events throughout Eurasia are examined in this study. Daily ERA5 precipitation and maximum temperature data (1991–2020) are analyzed at representative Eurasian locations. Droughts are quantified using SPI-3, heatwaves are identified using percentile-based temperature thresholds, and compound events are defined by their simultaneous occurrence. Statistical dependence and extreme co-occurrence are evaluated using probability metrics and copula-based modeling. Results reveal pronounced spatial heterogeneity in compound drought–heatwave characteristics across Eurasia. Mediterranean and western Eurasian locations exhibit frequent compound occurrences and strong upper-tail dependence, indicating synchronized extreme behavior. Continental interior regions show weak dependence, where compound events emerge primarily during severe drought years coinciding with persistent heatwaves. East Asian and monsoon-influenced regions display moderate compound activity driven by frequent heatwaves but weaker drought control. Extreme-event analysis further demonstrates that compound events amplify risk more strongly than heatwaves alone, with return periods indicating rare but severe joint extremes at drought-prone sites. Copula modeling confirms the existence of upper-tail dependence and emphasizes the importance of extreme-event synchronization rather than linear correlation by identifying the Gumbel family as the best match for the majority of locations. These results highlight the highly region-specific nature of compound drought-heatwave risk, which is controlled by underlying climate regimes. This study highlights the limitations of spatially averaged assessments and provides a statistically robust framework for diagnosing localized compound climate extremes. The findings encourage more focused climate risk assessment and adaptation planning throughout Eurasia and further knowledge of the effects of climate change.

ABSTRACT. Abstract River drainage pump stations are essential for mitigating inland flooding, but the pumps are typically activated only when inland water levels rise, leading to intermittent operation. This operating pattern creates long standby periods and short operating bursts, while true fault cases are rare and difficult to label. As a result, conventional supervised diagnostics are often impractical. This study develops a label-efficient anomaly detection model and a near-term risk forecasting scheme for drainage pump stations using unsupervised learning and interpretable health indicators. A compact sensing module installed on the pump records vibration and rotational-motion signals. The data are summarized over fixed time windows and normalized to reduce site-specific noise and operating variability. Three indicators are derived to represent key degradation signatures: overall vibration level, impulsive shocks, and rotational variability. An ensemble of unsupervised detectors learns the normal pattern only and flags deviations using majority voting with a persistence rule to suppress transient spikes. The detection outputs are integrated into an interpretable 0–100 risk score and translated into three operational levels (Green/Yellow/Red) for standard operating procedures. In addition, a lightweight interpretation layer suggests likely fault causes (e.g., bearing defect, unbalance, or looseness), and a sequence forecasting model estimates the likelihood of near-term risk escalation within a 2–3 h horizon to support proactive maintenance. The proposed framework is implemented and demonstrated using data from a controlled testbed and an operational pilot pump station, indicating practical feasibility for anomaly detection and early warning under intermittent operation. Acknowledgments This work was supported by Korea Environment Industry & Technology Institute (KEITI) through Project funded by Korea Ministry of Climate, Energy, Environment (MCEE) (Grant RS-2024-00331809).

ABSTRACT. Understanding phenological behaviour is increasingly important for advancing climate change mitigation, particularly in rapidly urbanising tropical regions. This study aims to advance climate mitigation frameworks by integrating phenological dynamics to assess urban–rural climate responses in Kuala Lumpur, Malaysia. The key challenge addressed is the limited incorporation of phenological signals into climate mitigation planning, despite their strong sensitivity to hydroclimatic variability and land-use change in tropical environments. To address this gap, a phenology-based analytical framework was developed using long-term satellite products and gridded climate data. Phenological indicators, including start of season (SOS), end of season (EOS), and growing season length (GSL), were derived to characterise phenological variation, while seasonal and annual hydro-meteorological variables were analysed to examine climatic drivers. Comparative phenology analyses were conducted to quantify urban–rural differences, and temporal trend analyses were employed to capture phenological dynamics and their responses to climate variability. The results reveal distinct urban–rural contrasts in phenological timing and dynamics across Malaysia, with urban areas exhibiting altered seasonal behaviour and reduced climate sensitivity compared to surrounding rural regions. These differences highlight the influence of urbanisation on phenological responses under tropical climate conditions. The findings demonstrate that phenological dynamics provide valuable indicators for understanding climate–land interactions and offer a robust basis for enhancing climate mitigation frameworks. Integrating phenological information into climate mitigation strategies can support more targeted, evidence-based planning for urban climate resilience and sustainable development in Kuala Lumour, Malaysia.

ABSTRACT. [Abstract]

Potential Evapotranspiration (PET) represents the maximum amount of evapotranspiration that can occur under given atmospheric conditions, assuming sufficient water availability at the land surface, and is closely related to drought occurrence and water resource scarcity. Accordingly, PET forecasting has been recognized as an important component for early drought response and efficient water resource management. In this study, a one-month-ahead PET forecasting framework was developed using latest artificial intelligence (AI) prediction models, and its applicability was evaluated. The forecasting models were designed to predict future PET based on historical PET time-series data. The study area covered the entire South Korean region, and PET data were derived from Terra Moderate resolution imaging spectroradiometer (MODIS) satellite products, enabling pixel-level one-month-ahead PET predictions. In addition to PET, various meteorological and topographic variables were considered as candidate input features for the models. Several time-series-based AI forecasting models were selected and applied under identical conditions, and their PET forecasting performances were comparatively analyzed. This study examines the applicability of PET forecasting by integrating satellite-based PET data with time-series AI prediction techniques, and the results are expected to provide fundamental insights for the development of more advanced drought forecasting and early response systems.

[Acknowledgment]

This work was supported by Korea Environment Industry & Technology Institute(KEITI) through Research and Development on the Technology for Securing the Water Resources Stability in Response to Future Change Project, funded by Korea Ministry of Climate, Energy, Environment(MCEE)(RS-2024-00332300).

ABSTRACT. Rising atmospheric CO₂ affects terrestrial hydrology through two primary mechanisms: radiative effects that modify the atmospheric energy balance and physiological effects that alter plant water use. However, the relative importance of these pathways varies across hydroclimatic regimes. This study quantifies the contributions of radiative (RAD) and physiological (PHY) CO₂ effects on streamflow in major Asian river basins using the Weather Research and Forecasting Hydrological Modeling System coupled with a glacier module (WRF-Hydro/Glacier). Meteorological forcing was derived from the CMIP6 1% CO₂ increase experiment (1pctCO₂), in which atmospheric CO₂ increases by 1% annually from pre-industrial levels. Factorial experiments were conducted using the 1pctCO₂-rad and 1pctCO₂-bgc configurations to separately isolate radiative and physiological forcings. Streamflow responses were analyzed for the Indus, Ganges, Yangtze, Yellow, Brahmaputra, and Mekong basins. Under elevated CO₂, streamflow increases by approximately 40–70% across all basins. Basin-scale aridity was then used to diagnose the dominant driving mechanisms. The Indus and Ganges basins were classified as arid, water-limited systems, while the Yangtze and Mekong basins were identified as humid, energy-limited systems. Results indicate that radiative effects dominate streamflow increases in arid basins, whereas physiological effects play a stronger role in humid basins by reducing transpiration and enhancing soil moisture retention. These findings highlight basin aridity as a key control on the relative importance of CO₂-driven radiative and physiological pathways, providing a mechanistic framework for understanding regional streamflow responses under climate change.

Acknowledgements This work was supported by Korea Environmental Industry & Technology Institute(KEITI) through Water Management Program for Drought, funded by Korea Ministry of Climate, Energy and Environment(MCEE)(RS-2023-00231944).

ABSTRACT. Coastal hazards associated with tropical cyclones (TCs) have historically received limited attention in Indonesia due to its equatorial location, where TC genesis and sustained tracks are climatologically rare. Consequently, storm surge hazards have long been considered secondary compared to seismic-driven threats. However, recent events challenge this assumption. Tropical Cyclone Seroja in 2021 caused widespread coastal inundation and socioeconomic losses across eastern Indonesia, while Tropical Cyclone Senyar in 2025 developed anomalously close to the equator and affected northern Sumatra. These events underscore growing uncertainty in future TC behavior and expose a critical knowledge gap regarding the absence of quantitative, climate-informed projections of TC-induced storm surge hazards in Indonesia.

This study aims to assess future storm surge hazards associated with tropical cyclones in Eastern Indonesia under climate change conditions. To address the limited historical record, synthetic typhoons were generated using a statistical approach that integrates representative historical storm characteristics with projected changes in intensity and track patterns. Future cyclonic dynamics were derived under the high-emission scenario RCP 8.5 using an ensemble of four global climate models: HadGEM, CMCC, CNRM-CM6, and EC-EARTH. A high-resolution hydrodynamic model was then applied to simulate storm surge generation, propagation, and coastal runup across the complex archipelagic setting of Eastern Indonesia.

Results indicate a substantial increase in extreme storm surge heights driven by both intensified TC wind fields and projected shifts in storm track trajectories. The amplification of surge and wave conditions significantly elevates coastal flooding potential, particularly in Kupang (East Nusa Tenggara), identified as a regional hotspot due to its exposure and geomorphological setting. The findings demonstrate that, although TCs remain relatively infrequent near the equator, their projected intensification under climate change may lead to disproportionately severe coastal impacts.

The interaction between projected changes in tropical cyclone intensity, track variability, and Indonesia’s complex archipelagic bathymetry has not been systematically evaluated, limiting the development of robust coastal adaptation and design standards. This study provides the first climate-informed, region-specific projection of TC-induced storm surge hazards for Eastern Indonesia. The results establish a scientific basis for updating coastal design standards, disaster mitigation strategies, and long-term adaptation planning, contributing to enhanced resilience of vulnerable coastal communities in a warming climate.

ABSTRACT. In recent years, the growing impacts of climate change on rivers have made it increasingly important to examine streamflow trends for effective water resource planning and management, particularly in arid and semi-arid regions with diverse geography and pronounced climatic variability. This study investigates how streamflow has changed over the last decades across Kazakhstan, Central Asia. Monthly, seasonal, and annual streamflow trends at 30 river stations were analyzed, representing four major basin groups in western, northern, eastern, and south-eastern regions of the country for the period 1993-2023. Three non-parametric trend tests based on the Mann-Kendall framework (Mann–Kendall, Modified Mann–Kendall, and Block-Bootstrapped Mann–Kendall) were applied to detect statistically significant trends at the 5% level. At the annual scale, most rivers do not exhibit statistically significant trends. However, seasonal and monthly analyses revealed greater and more spatially structured changes. Western rivers showed widespread declines in winter, summer, and autumn flows, while spring snowmelt flows decreased across multiple basins in all regions. Many rivers in eastern and northern Kazakhstan show significant winter flow increases, suggesting enhanced cold-season runoff and greater contribution of snowmelt outside the traditional spring melt peak. Monthly trends are mainly consistent with the seasonal patterns, which increases confidence in the robustness of the detected signals. Overall, the findings indicate that more arid lowland basins in Central Asia may be experiencing faster and more substantial reductions in streamflow than mountainous semi-arid basins. These emerging patterns have important implications for regional water management, ecosystem sustainability, and drought preparedness under a warming climate.

Acknowledgement: This work was supported partially by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. NRF-RS-2023-00209531). This work is financially supported by Korea Ministry of Climate, Energy, Environment(MCEE) as 「Graduate School specialized in Climate Change」

ABSTRACT. Climate change is increasing hydrologic variability in river systems, resulting in more frequent and persistent turbidity and fine sediment exposure driven by extreme precipitation, altered flow regimes, and watershed-scale disturbances. Despite growing recognition of sediment-related ecological degradation under climate change, most ecological assessments continue to rely on short-term concentration-based metrics that inadequately capture cumulative and time-lagged biological responses. This study presents an integrated framework to quantify and predict riverine biotic responses to cumulative turbidity stress using sensitivity-based biological indices and generalized additive models (GAMs).

Long-term suspended sediment (SS) exposure was quantified as cumulative dose (time-integrated SS concentration, mg·hour/L) over multiple antecedent periods (1, 3, and 6 months), incorporating threshold-exceedance metrics that characterize both the duration and magnitude of exceedance beyond ecologically meaningful SS levels. To represent biological susceptibility to chronic sediment stress, a novel Turbidity Sensitivity Index for benthic macroinvertebrates (TSI-BM) was developed by weighting taxa according to sensitivity classes derived from consistent monotonic response relationships observed in Ephemeroptera, Plecoptera, and Trichoptera (EPT) assemblages. The framework was applied to 22 long-term monitoring sites in the upper North Han River Basin, Republic of Korea, a river system increasingly affected by climate-driven sediment dynamics.

Results indicated that TSI-BM exhibited a statistically significant and robust negative association with six-month cumulative turbidity exposure (Spearman’s ρ = −0.46, p < 0.001), demonstrating superior explanatory performance relative to conventional macroinvertebrate indices based on instantaneous water-quality metrics. GAM analyses further identified pronounced nonlinear effects and interaction structures among cumulative turbidity stress, hydraulic variables (water depth and flow velocity), and nutrient concentrations, with cross-validated coefficients of determination evaluated at individual monitoring sites ranging from 0.40 to 0.72. These findings underscore the importance of jointly accounting for climate-induced alterations in sediment regimes and habitat conditions when assessing ecological sensitivity.

Overall, this study demonstrates that integrating cumulative turbidity exposure with sensitivity-weighted biological indices and flexible nonlinear modeling constitutes a methodologically robust and broadly applicable analytical framework for evaluating climate change impacts on riverine ecosystems. The proposed approach supports climate-adaptive river monitoring, sediment impact assessment, and evidence-based sediment management under increasingly variable hydrologic conditions.

This work was supported by the Korea Environmental Industry and Technology Institute (KEITI) through Aquatic Ecosystem Conservation Research Program, funded by Korea Ministry of Environment (MOE) (RS-2021-KE001374).

ABSTRACT. Wetlands play a pivotal role in the global carbon cycle, acting as both significant carbon sinks and potential sources of greenhouse gases. In the context of accelerating global climate change, understanding the carbon budget of riverine wetlands is essential for developing effective national adaptation and mitigation strategies. This study focuses on evaluating the impact of climate change on the future carbon cycle within Korean riverine wetlands, which represent critical ecological zones with distinct hydrological and climatic characteristics. To achieve high-precision simulations, the Community Land Model (CLM) was employed as the core modeling framework. Spatial meteorological forcing data were synthesized using PRISM and LDAPS to reflect local climate variability accurately. A robust parameter optimization process was implemented using Latin Hypercube Sampling (LHS) to generate 1,000 parameter sets, followed by a Generalized Likelihood Uncertainty Estimation (GLUE) analysis. Through this framework, optimal parameters were derived to maximize consistency with observation data, and the variability of carbon sinks and emissions in Korean riverine wetlands was quantified by projecting future climate change scenarios onto the optimized model. The integration of high-resolution forcing data and the GLUE-based optimization framework significantly enhances the model's capacity to simulate complex biogeochemical processes in riverine environments. The study identifies the sensitivity of carbon fluxes to changing climatic variables, revealing critical shifts in carbon exchange mechanisms under future climate projections. Ultimately, these findings provide a quantitative scientific basis for establishing data-driven carbon management and climate change adaptation strategies for riverine wetlands.

Acknowledgement: This work was supported by the National Research Foundation of Korea, which was funded by the Ministry of Science, ICT and Future Planning (RS-2024-00457308), and Korea Environment Industry & Technology Institute (KEITI) through ‘Wetland Ecosystem Value Evaluation and Carbon Absorption Value Promotion Technology Development Project’, funded by Korea Ministry of Climate, Energy and Environment(MCEE) (RS-2022-KE002030)

ABSTRACT. Against the backdrop of intensifying global climate change, regional assessments are crucial for developing adaptation strategies. The Qinhuai River Basin (QRB), a tributary of the lower Yangtze River in China, faces heightened flood risks and water resource stress due to climate change and rapid urbanization. This study aims to bridge a gap in high-resolution climate projections by developing a modeling framework to assess the impacts of climate change in the QRB. The objectives include (1) developing a semi-distributed hydrological model (HEC-HMS) to simulate extreme flood events; (2) applying bias correction techniques to downscale CMIP6 global climate model projections; and (3) analyzing historical and future trends in temperature and precipitation under various socio-economic pathways (SSPs).

This research confronts several challenges, including increasing flood-disaster severity, the inadequacy of coarse-resolution global climate models for local studies, biases in raw climate outputs, and the lack of studies that combine hydrological modeling with CMIP6 projections for flood risk assessment. The methodology involves building an HEC-HMS model for the QRB using high-resolution data for watershed delineation and parameterization, calibrating it against streamflow data, and validating it using performance metrics such as Nash-Sutcliffe Efficiency (NSE), Percent Bias (PBIAS), and the Coefficient of Determination (R2). Subsequently, the study employs the CMhyd tool to bias-correct climate projections using observed temperature and precipitation data from 1980 to 2020. It evaluates several bias-correction methods and generates corrected data for three SSP scenarios (SSP1-2.6, SSP2-4.5, and SSP5-8.5) that project future climate. Key findings reveal consistent warming trends, with significant temperature increases and substantial variability in projected precipitation, indicating an intensified hydrological cycle and more frequent flooding.

The study concludes that the integrated framework establishes a scientific basis for proactive flood risk management and adaptive water resource planning in the QRB. The urgency of implementing climate mitigation and adaptation strategies is emphasized, including the need for updated infrastructure and the integration of climate projections into long-term governance to enhance resilience to climate change.

ABSTRACT. Understanding regional variability in extreme snowfall is essential for hydrological design and disaster risk reduction, especially under changing climate conditions. This study develops a robust methodology to classify and interpret spatial patterns of heavy snow accumulation across South Korea. Maximum annual snow depth data from 77 stations were analyzed using five statistical distributions to estimate return levels over various time intervals. Distribution fitting was validated through statistical goodness-of-fit tests to capture localized tail behavior. The spatial extent of extreme snow impact was interpolated using geostatistical techniques, providing detailed regional maps. To segment homogeneous zones, unsupervised learning methods including Gaussian Mixture Models, K-means, and Hierarchical Clustering were applied. Clustering robustness was examined using bootstrap resampling and Adjusted Rand Index (ARI), where GMM outperformed other methods in spatial stability. These results offer critical insights into regional snow risk zoning, contributing to infrastructure resilience planning and adaptive water resources management in mountainous and coastal regions affected by snowfall extremes.

ABSTRACT. The Kyushu region of Japan is one of the areas where the occurrence of linear rainfall bands is notably high compared to the rest of Japan. It is pointed out that future climate warming will lead to more frequent and more severe torrential rains. According to the IPCC Fifth Assessment Report, warming could reach up to 4.8°C by 2100. Using a Self-Organizing Map (SOM), we divided the current climate meteorological fields (east-west wind speed, north-south wind speed, and precipitable water) from 2006 to 2023 into 480 units and 40 groups. We then extracted linear rainfall bands from JMA's analyzed rainfall data and linked them to the meteorological field classifications by SOM, identifying meteorological field groups prone to linear rainfall bands formation. The results revealed that in Kyushu region, the number of linear rainfall bands occurring was high in meteorological field groups characterized by the inflow of moist air from East China Sea. The top three groups were G34, G30, and G33. Next, meteorological fields were extracted from past experiments, 2°C warming experiments, and 4°C warming experiments of the MIROC5 model in d4PDF (database for Policy Decision-making for Future climate). Using SOM, classification was performed to analyze changes in the occurrence frequency of meteorological field groups prone to linear rainfall bands. The results showed that in meteorological field G34, which frequently produces linear rainfall bands in southern Kyushu, the occurrence frequency increased due to rising temperatures. In G30, where linear rainfall band cases were sporadically observed across all of Kyushu, the frequency decreased with rising temperatures. In G33, which frequently produced linear rainfall bands in northern Kyushu, the frequency increased in the 2°C warming experiment compared to the past experiment but decreased in the 4°C warming experiment compared to the 2°C warming experiment. Among other meteorological field groups with linear rainfall bands cases in the Kyushu region, some groups showed increased occurrence frequency with warming, while others showed decreased frequency. However, when combining the top 10 groups with the highest number of linear rainfall bands occurrence cases and comparing the meteorological field occurrence frequency, a clear increase due to warming became evident. This result suggests that heavy rainfall accompanied by linear rainfall bands will increase in Japan's Kyushu region due to global warming.

ABSTRACT. Climate change intensifies thermal stratification, degrading lake water quality and ecosystems. In Lake Suwa, rising temperatures threaten to expand hypoxic water masses. While accurate numerical simulations are vital for lake management, sparse observation points and coarse meteorological datasets often fail to capture the spatial heterogeneity of lake winds—the primary driver of water circulation. This study addresses this gap by coupling a high-resolution atmospheric model with a hydrodynamic water quality model. We integrated a meteorological model (SCALE, Nishizawa et al. 2015, Sato et al. 2015) with an aquatic ecosystem model (AEM3D, www.hydronumerics.com.au) to simulate three scenarios: Case 1 (with observed meteorological data), Case 2 (present conditions via SCALE), and Case 3 (future projections via pseudo-warming data made with SCALE output). Comparing Case 1 and Case 2 verified the atmospheric model's effectiveness, while the Case 2–Case 3 comparison revealed future trends. The SCALE model successfully reproduced diurnal wind cycles. Although the model required bias correction for temperature, it predicted uniform seasonal warming with minimal changes in wind speed, reflecting the region's topographically driven climate. Case 2 captured realistic horizontal surface circulations that point-based observations in Case 1 missed, despite a slight decrease in DO accuracy at the lake's center. Future projections (Case 3) show that higher temperatures accelerate phytoplankton growth and deplete surface DO during peak heat. Most critically, the hypoxic area (DO < 3 mg/L) expands significantly from May to September. We project that the duration of bottom-layer hypoxia will increase by an average of 17.9 days lake-wide, and by 30.6 days in the deepest basins. High-resolution atmospheric coupling provides a more realistic representation of lake currents than traditional interpolation. Our findings confirm that climate change will substantially prolong and expand hypoxia in Lake Suwa, posing a severe challenge for future water quality management.

ABSTRACT. This study quantitatively evaluates the impacts of climate change on meteorological and hydrological components in the Nakdong River basin, South Korea, by coupling multiple GCM-based SSP scenarios with the watershed model SWAT (Soil and Water Assessment Tool). To account for climate model uncertainty, we used multiple GCM-based SSP scenarios. Bias-corrected and downscaled daily precipitation, temperature, wind speed, relative humidity, and solar radiation for historical and future periods were provided as inputs to SWAT to analyze annual and seasonal changes in key hydrological components such as runoff, evapotranspiration, and groundwater recharge. The results show that precipitation, evapotranspiration, runoff, and recharge generally increase toward the late-century period, with relatively larger increases under the SSP5-8.5 scenario. From a water balance perspective, both SSP2-4.5 and SSP5-8.5 scenarios indicate decreasing evapotranspiration ratios relative to precipitation and increasing runoff ratios, whereas the recharge ratio remains relatively stable despite increasing absolute recharge amounts toward the future. Thus, these findings suggest that climate-induced increases in total water availability and enhanced runoff proportions may substantially influence long-term water supply planning, flood and drought risk management, and reservoir operation strategies in the Nakdong River basin.

Funding: This work was carried out under the KICT Research Program (Development of IWRM Korea Technical Convergence Platform Based on Digital New Deal, Grant number 20260156-001) funded by the Ministry of Science and ICT.

ABSTRACT. Understanding how extreme rainfall frequency characteristics evolve over time is essential for hydrologic safety, flood risk mitigation, and resilient infrastructure planning in monsoon-dominated regions such as South Korea. Traditional rainfall frequency analysis commonly assumes stationarity, which may no longer be valid under observed changes in extreme precipitation. This study investigates trend-driven nonstationarity in extreme rainfall and evaluates alternative nonstationary frequency models using nationwide meteorological observations. Annual maximum precipitation (AMP) series across multiple rainfall durations, from sub-daily to multi-day events, are first analyzed using non-parametric trend detection methods to identify statistically significant monotonic trends. Stations exhibiting significant increasing trends are then selected to ensure that nonstationary modeling is supported by observational evidence. For these stations, correlation analysis is conducted to examine statistical associations between extreme rainfall, time (year), and global mean temperature, providing a basis for formulating alternative nonstationary models. A suite of stationary and nonstationary Generalized Extreme Value (GEV) models is subsequently developed, allowing the location and scale parameters to vary with time or global mean temperature. These variables are treated as indexing factors that represent the gradual evolution of rainfall frequency characteristics rather than explicit physical drivers. Model performance is systematically compared using information-based criteria and likelihood diagnostics to identify the most suitable model structures. The results indicate that stations with significant increasing trends exhibit pronounced nonstationary behavior, particularly for short-duration rainfall extremes. Nonstationary GEV models generally outperform stationary formulations and yield higher and more variable return level estimates for design-relevant return periods, highlighting the sensitivity of design rainfall to model choice. Overall, the findings demonstrate that a trend-informed, model-comparative nonstationary framework provides a robust and practical basis for improving rainfall frequency estimation and enhancing resilience against flood hazards in South Korea.

Acknowledgement 'This work was supported partially by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. NRF-RS-2023-00209531) his work is financially supported by Korea Ministry of Climate, Energy, Environment (MCEE) as 「Graduate School specialized in Climate Change」 '

ABSTRACT. Climate change driven by anthropogenic greenhouse gas emissions has altered the global hydrological cycle, increasing the frequency and intensity of extreme weather events such as heavy rainfall and floods. These changes significantly affect terrestrial ecosystems and human activities, particularly by intensifying soil erosion processes. Among various erosion mechanisms, water-induced soil erosion plays a critical role in redistributing soil particles and mobilizing soil organic carbon (SOC), which represents one of the largest carbon pools in terrestrial ecosystems. During erosion processes, SOC stored in soils can be transported and decomposed, leading to the release of carbon dioxide into the atmosphere and generating positive feedback on climate change. However, the interactions among hydrological processes, soil erosion, and SOC dynamics remain highly uncertain, especially under future climate conditions. Many previous studies have relied on empirical models or simplified assumptions, limiting their ability to realistically represent erosion-driven carbon fluxes. This study applies a physically based hydrological model to explicitly simulate soil erosion processes and associated SOC dynamics at the watershed scale. The model integrates surface runoff generation, sediment transport, and carbon redistribution to quantify SOC loss, transport, and potential deposition. By coupling hydrological and erosion processes with carbon dynamics, the model provides a mechanistic framework for assessing the impacts of changing precipitation patterns and hydrologic extremes on SOC behavior. Climate change scenarios are incorporated to evaluate future variations in soil erosion and SOC dynamics. Scenario-based simulations enable the analysis of spatiotemporal changes in SOC and the identification of regions and periods particularly vulnerable to erosion-induced carbon losses. The results of this study are expected to enhance the understanding of erosion–carbon–climate feedback mechanisms and to support sustainable land and watershed management strategies aimed at mitigating soil carbon losses under future climate change.

Funding: This work was supported by Korea Environmental Industry&Technology Institute(KEITI) through R&D Program for Innovative Flood Protection Technologies against Climate Crisis Project, funded by Korea Ministry of Environment(MOE)(2022003460002)

ABSTRACT. The purpose of the work Vegetation dynamics are fundamental components of terrestrial ecosystems that regulate the exchange of water and energy. However, the hydrological feedback of the observed "global greening" remains complex. The primary purpose of this work is to clarify the ambiguous impact of vegetation growth on soil moisture (SM), specifically addressing whether greening leads to soil drying or wetting on a global scale.

Key issues addressed A critical gap exists in understanding the spatial heterogeneity of vegetation-water coupling. While greening is generally viewed as beneficial for carbon sequestration, its potential to exacerbate soil drought through enhanced evapotranspiration is debated. Previous studies have yielded conflicting results due to limited regional representations and discrepancies in model parameterizations, creating uncertainty for future ecological planning.

The methodology used To address these challenges, this study integrated a robust dataset comprising multi-source satellite observations, reanalysis products, and outputs from 12 Earth System Models (ESMs) from the CMIP6 archive. We standardized these datasets to a unified 0.25° resolution. The analysis covered a historical period (1982–2020) to quantify past coupling trends and extended to the future (2015–2100) under shared socioeconomic pathways (SSPs). Advanced statistical methods, including Granger causality tests and sensitivity analyses, were employed to disentangle the dominant drivers among precipitation, evapotranspiration, and vegetation indices.

Results and conclusions. Approximately 49.96% of global vegetated areas exhibit a "greening-drying" pattern. This phenomenon is predominantly attributed to excessive vegetation transpiration, with sensitivity analyses indicating that transpiration explains 42% to 82% of the SM variance, particularly in grasslands and cultivated lands. Furthermore, ESM projections indicate that this vegetation-induced soil dryness is likely to persist and intensify in the future. We conclude that ignoring the soil moisture carrying capacity in ecological restoration could aggravate soil drought risks. These findings highlight the critical need to balance ecosystem carbon goals with water security in climate change adaptation strategies.

ABSTRACT. Hydrological connectivity is crucial for understanding water-ecosystem dynamics, as it serves as a key link between different landscape units. However, the variability of hydrological connectivity in Central Asia remains unexplored, which poses challenges to a comprehensive understanding of ecohydrological processes. This study investigates the spatiotemporal patterns and driving mechanisms of hydrological connectivity in the Tarim River Basin (TRB), Central Asia, from 1990 to 2020, employing a novel approach that integrates remote sensing and reanalysis data. The results indicate an increasing trend in the hydrological connectivity index (HCI), with approximately 60% of the TRB exhibiting significant increases. Climate change exerts the greatest direct (0.59) and total (0.64) effects on HCI, with potential evapotranspiration (19.2%) and temperature (12.6%) being the dominant factors. In mountainous regions, climate change (0.65) is the primary driver, while human activities have a greater impact in the plains (−0.27). These findings offer a new framework for studying ecohydrological processes in arid regions.

ABSTRACT. Recently climate change and abnormal weather conditions occurs such as the increase of heat waves period, droughts, amount of sunlight, multiple torrential rains and decreasing the rainy season. It is pointed out as a problem that not COD increase, water pollution, the limits of the BOD improvement, algae growth and death of fish according to the continuing inflow of non-point sources, but also changes in river environment according to the various river projects such as the increase of river depth and retention time, decrease of flow rate, and pollutant accumulation. For this reason proactive water quality improvement and management skills are needed for field conditions. Recently micro-bubble applied DAF processes are attracting the interest for algae and suspended solids control in stagnant water areas. In DAF process, the main mechanism governing the particle removal is the collision between particle and bubble (AWWA, 1999), and it is understood that the effect increases when the diameters of contaminant particles and bubbles are in similar ranges (Han, 2001). Since the number, diameter and surface area of micro bubbles are dependent on the pressure and flow changes during the generation, it is necessary to deduce appropriate pressure and flow level for its operation. This study aimed to draw a most fit operational bubble generating properties by monitoring bubble numbers, size and surface changes under various operational pressures and flow.

ABSTRACT. Regional Frequency Analysis (RFA) estimates stable probabilistic hydrological quantities by defining homogeneous regions in areas with short observation records, such as South Korea. In South Korea, rainfall observation stations are currently grouped into 26 homogeneous regions, and the data are jointly pooled; however, this regional delineation is based on rainfall data collected up to 2017. Recent climate-driven changes in rainfall characteristics suggest that the adequacy of the existing regional delineation should be reassessed using updated data. Violations of the assumptions regarding predefined regions have been observed through a reassessment based on recent rainfall data. Therefore, this study aims to delineate homogeneous regions for rainfall regional frequency analysis based on recent rainfall data using advanced clustering analysis techniques. The study compares two methodological approaches: 1) standard Euclidean distance-based K-means clustering, and 2) K-means clustering applied after dimensionality reduction using t-distributed Stochastic Neighbor Embedding(t-SNE), which preserves similarity in high-dimensional data within a low-dimensional space. To identify the temporal evolution of homogeneous regions due to climate change, these approaches are applied to cumulative rainfall datasets up to 2017, 2021, and 2025, respectively. The reclassified results will be compared and evaluated by clustering method using heterogeneity measures. Through this study, it is expected that the influence of ongoing climate change on homogeneous region delineation in rainfall regional frequency analysis can be identified. The results of this study can be used as fundamental data for future homogeneous region analysis in South Korea.

ABSTRACT. Recent climate change, characterized by rising temperatures and increasingly irregular precipitation, has intensified drought risk and its impacts on agriculture. Meteorological drought indices such as the Standardized Precipitation Index (SPI) and the Standardized Precipitation Evapotranspiration Index (SPEI) are widely used to describe drought conditions, yet their signals do not always correspond closely with observed agricultural impacts. Impacts may occur earlier or later than index variations, and the relationships differ by indicator type and temporal scale. This study investigates lead–lag relationships between drought indices and observed agricultural impact indicators, while also incorporating annual emergency response records available only at yearly resolution. Monthly agricultural indicators include paddy water-shortage area, field-crop wilting area, and a soil-moisture-based field drought index (SSI-3) derived from soil water availability ratio data. SPI and SPEI are calculated at accumulation scales from 1 to 36 months to represent both short-term and cumulative drought conditions. For monthly indicators, correlation analyses are conducted with systematically shifted time series to identify the lag and accumulation scale yielding the most interpretable index–impact relationships. To stabilize interpretation across heterogeneous indicators, results are synthesized using a role-based framework distinguishing fast-response indicators reflecting near-term crop stress from cumulative-impact indicators representing prolonged moisture deficits. Annual emergency response measures—including emergency water-source development, equipment support, and manpower mobilization—are analyzed using same-year correlations with drought indices aggregated to annual metrics. Because these variables are reported annually, lead–lag estimation is not applied; instead, their associations are evaluated in a complementary manner to assess how response activities co-vary with drought severity. Results reveal clear timing differences between index signals and observed monthly impacts, with shorter accumulation scales associated with more immediate responses and longer scales linked to prolonged impact conditions. Annual response records show consistent co-variation with drought severity in several cases, supporting their use as supplementary evidence of drought-related management burden. Overall, combining lead–lag analysis for monthly impacts with annual association analysis for response measures enhances impact-relevant drought interpretation and informs practical lead times for agricultural drought early warning. This work was supported by Korea Environment Industry & Technology Institute(KEITI) through Climate Resilient R&D Project for Water-Related Disaster Management, funded by Korea Ministry of Climate, Energy and Environment(MCEE)(RS-2023-00230286)

ABSTRACT. This study assesses projected changes in extreme flood risk within the Miho River Basin under future climate change scenarios, employing a multi-model ensemble of 23 CMIP6 global climate models (GCMs). The primary objective is to quantify variations in rainfall and flood quantiles across distinct future periods, thereby providing a scientific basis for designing climate-resilient flood infrastructure in South Korea. Bias-corrected daily precipitation data were obtained from CMIP6 GCMs under the SSP2-4.5 and SSP5-8.5 scenarios. From these data, annual maximum daily rainfall series were derived and fitted to the Generalized Extreme Value (GEV) distribution to estimate rainfall quantiles. The analysis encompasses the historical baseline period (1985–2014) and three future intervals: near-term (2015–2044), mid-term (2043–2072), and long-term (2071–2100). Changes in flood quantiles were inferred based on established hydrological relationships between rainfall inputs and corresponding flood responses, referencing prior event-based studies conducted in the Miho River Basin. The results indicate a significant intensification of rainfall quantiles and the associated flood quantiles across all future periods. These projected increases align with the nonlinear amplification effects commonly observed in rainfall–runoff dynamics under a warming climate. Spatial patterns suggest that the upper tributaries and inland regions of the Miho River Basin exhibit greater sensitivity to extreme precipitation changes. Inter-model uncertainty was rigorously assessed, and ensemble-based statistics were applied to derive robust and representative future trends. The findings underscore the importance of updating current flood quantile standards and reevaluating long-term flood risk management strategies within the basin. Moreover, the framework developed in this pilot study offers a scalable and transferable methodology for climate change impact assessment in other major river basins across South Korea.

ABSTRACT. Recent climate change has increased the frequency and intensity of droughts, resulting in escalating damages to the agricultural sector. Rising air temperatures and increasingly irregular precipitation patterns have intensified pressure on agricultural water resources and exacerbated drought impacts on crop production and water management systems. Because agricultural drought is an impact-oriented phenomenon characterized by vegetation stress and yield loss rather than solely by precipitation deficits, quantitative assessment from agricultural, hydrological, environmental, and economic perspectives is essential for effective drought response and management. Conventional meteorological drought indices such as the Standardized Precipitation Index (SPI) rely solely on precipitation and thus have limitations in directly capturing vegetation responses. This study employs satellite remote sensing data to quantitatively assess agricultural drought impacts using the Drought Severity Index (DSI), derived from Moderate Resolution Imaging Spectroradiometer (MODIS) evapotranspiration (ET) and potential evapotranspiration (PET) products. DSI enables direct representation of vegetation water stress and ecohydrological responses that are difficult to characterize using precipitation or soil moisture alone. DSI performance was evaluated through comparison with the Normalized Difference Vegetation Index (NDVI), and its relationship with crop yield anomalies relative to long-term normals was analyzed to quantify agricultural drought impacts. The results demonstrate that satellite-based drought indices provide robust and spatially explicit information for agricultural drought impact assessment and can serve as critical baseline data for developing satellite-based drought monitoring systems and agricultural drought response strategies.

This work was supported by Korea Environment Industry & Technology Institute(KEITI) through Climate Resilient R&D Project for Water-Related Disaster Management, funded by Korea Ministry of Climate, Energy and Environment(MCEE)(RS-2023-00230286)

ABSTRACT. Accurate Climate change–induced increases in abnormal rainfall have intensified flood damage to farmland in Korea. Although drainage improvement projects have been implemented to reduce such damage, their effectiveness remains limited due to insufficient consideration of climate change impacts, particularly in conventional flood analysis techniques. Existing approaches rely primarily on hydrological analyses, which restrict their ability to realistically reproduce inundation processes in farmland.

This study proposes an advanced farmland flood analysis method that more accurately reproduces runoff and overflow phenomena in drainage channels under heavy rainfall conditions. The conventional GATE model is based on a water balance approach using available storage and watershed discharge. While this method enables rapid and simple analysis, it fails to reflect channel characteristics, drainage networks, hydraulic behavior, and unsteady flow conditions, leading to limited accuracy in inundation prediction.

To overcome these limitations, an improved flood analysis framework incorporating channel networks, hydraulic characteristics, and unsteady flow was developed using the SWMM dynamic wave model. The proposed methodology was applied to the JB district of Cheongju-si, Chungcheongbuk-do, where severe flooding occurred during an extreme rainfall event in July 2017, with a total rainfall of 290.2 mm and a maximum intensity of 91.8 mm/hr. Simulation results from the advanced model were compared with those from the GATE model and validated against observed inundation depths.

The results showed that the SWMM dynamic wave model more accurately reproduced observed flood levels and inundation occurrence, particularly at critical locations where the GATE model failed to simulate flooding. These findings indicate that the proposed approach provides a more reliable representation of farmland inundation and can contribute to more effective planning and design of drainage improvement measures under changing climate conditions.

Acknowledgments: This work was supported by Korea Institute of Planning and Evaluation for Technology in Food, Agriculture and Forestry(IPET) through Intelligent Agricultural Infra Management for Climate Change Development Program, funded by Ministry of Agriculture, Food and Rural Affairs(MAFRA)(RS-2025-02303335)

ABSTRACT. Climate change threatens global water security by altering hydrological regimes and increasing supply uncertainty. This study evaluates the effectiveness of a national adaptation plan in shaping future water security by applying an integrated modeling framework that couples the Soil and Water Assessment Tool (SWAT) and the Water Evaluation and Planning (WEAP) system to the Seomjin River Basin in South Korea. The research identifies how policy interventions modify basin-scale system performance across domestic, industrial, and agricultural sectors under diverse Shared Socioeconomic Pathway (SSP) climate scenarios. Principal results demonstrate that the national plan significantly stabilizes domestic and industrial water supplies. In these sectors, reliability and resilience indices remained high, and vulnerability decreased by approximately 16% to 100% due to infrastructure expansion and multi-source integration. Conversely, agricultural water and instream flows showed limited responsiveness to adaptation, remaining heavily governed by natural hydrological variability. While the strategy enhanced apparent supply stability, the Water Footprint Self-Sufficiency (WFSS) analysis revealed that structural dependence on external water resources persisted or slightly increased in key industrial hubs. Major conclusions indicate that infrastructure-based adaptation can mask latent structural vulnerabilities. Future adaptation planning must shift from uniform supply-oriented measures toward sector-specific, hydrologically grounded strategies to ensure sustainable basin management under accelerated climate change

ABSTRACT. The purpose of this study is to improve long-term water supply forecasting in South Korea under increasing drought risk caused by climate change. For this purpose, an Ensemble Streamflow Prediction (ESP)-based inflow forecasting program was developed. The program integrates long-range probabilistic climate forecasts from the Korea Meteorological Administration with historical hydrological data to generate monthly dam inflow scenarios for major river basins. These scenarios are used to support drought preparedness and water-resources management decisions. Traditional long-term inflow forecasting methods are mainly based on historical average inflows. However, this approach has clear limitations because it cannot reflect year-to-year climate variability or extreme weather conditions. During recent drought periods, such as 2022-2023, forecasts based on historical averages failed to capture actual hydrological changes, leading to reduced prediction accuracy. To overcome this limitation, this study adopted an ESP-based long-term forecasting approach that dynamically reflects current hydrological conditions. The forecasting program was developed through a two-step procedure. First, ensemble inflow scenarios were generated using the GR4J rainfall?runoff model with a snowmelt module, driven by daily meteorological data from 1966 to 2024. Precipitation events were classified into three categories based on basin-average rainfall. Probabilities from the long-range climate forecasts were then applied as weighting factors to estimate future inflows. The program was implemented in FORTRAN and fully automated, covering data processing, simulation, and post-processing. It was designed to be flexibly applied to mid-scale basins. The program was applied to four multi-purpose dams in the Geum River basin: Daecheong, Yongdam, Buan, and Boryeong dams. The results show that ESP-based forecasts generally performed better than traditional average-based forecasts in most seasons. In particular, the performance for Daecheong Dam improved significantly, with a 24.8% reduction in mean absolute error (MAE) and a 16.2% reduction in root mean square error (RMSE). However, for Boryeong Dam, the ESP performance was relatively lower due to basin characteristics and higher forecasting uncertainty. In conclusion, the ESP-based long-term inflow forecasting framework provides more realistic and reliable predictions than static statistical methods, especially under extreme climate conditions. The results indicate that the proposed framework is well suited for operational applications, such as drought early warning systems.

This work was supported by Korea Environmental Industry & Technology Institute (KEITI) through Water Management Program for Drought, funded by Korea Ministry of Climate, Energy and Environment (MCEE) (RS-2023-00231944).

ABSTRACT. Climate change is intensifying hydroclimatic extremes in rapidly urbanising tropical basins, demanding integrated approaches that bridge physical risk assessment and policy-driven adaptation. This study develops an integrated hydroclimate risk–governance assessment for the Klang River Basin, Malaysia, quantifying future flood dynamics under intermediate and high forcing pathways (RCP4.5 and RCP8.5) and evaluating the readiness of Nature-based Solutions (NbS) within national adaptation and disaster risk reduction systems. Calibrated rainfall–runoff and hydraulic modelling were conducted for four major sub-basins under projected climates of 2021–2050 and 2051–2080 time horizons. Under RCP4.5, results indicate increased discharge variability and moderate amplification of peak flows, with expansion of 100- and 200-year Average Recurrence Interval (ARI) flood extents in urban corridors. Under RCP8.5, hydrological intensification becomes substantially more pronounced, with systematic increases in extreme discharge and wider spatial growth of high-depth inundation zones (>1.5 m). Spatial integration of flood projections with demographic and socio-economic indicators reveals significant growth in exposed populations under both pathways, with disproportionately higher escalation in high and very high risk categories under RCP8.5. These findings suggest that existing structural mitigation measures alone are unlikely to offset escalating hydroclimatic pressures under high-emission futures. A policy integration assessment of NbS across climate, water, urban planning and disaster risk reduction frameworks indicates growing strategic recognition of ecosystem-based adaptation, yet persistent operational gaps in institutional coordination, regulatory enforceability, financing and performance monitoring. The divergence between projected flood intensification and current NbS mainstreaming highlights an emerging adaptation deficit. By coupling hydroclimate projections with institutional diagnostics, this study provides a translational framework linking physical risk escalation with governance reform, underscoring the urgency of embedding NbS within river basin management and risk-informed development planning in rapidly urbanising tropical regions.

ABSTRACT. Climate change is increasing the occurrence of hydrologic extremes, including severe droughts. This study combines the Weather Research and Forecasting (WRF) model with a Long Short-Term Memory (LSTM) network to generate and propagate physically conditioned extreme-drought scenarios in a Korean catchment. WRF is first used to produce weather series representing plausible future drought conditions, which are then fed into the LSTM to simulate corresponding runoff. Hydrological drought is quantified using the standardized streamflow index (SSFI) computed from the simulated flows. A sensitivity analysis of meteorological inputs shows that using only mean temperature performs similarly to using maximum and minimum temperature plus sunshine duration, suggesting that adding highly correlated variables can increase input uncertainty without improving runoff performance. For the 2015 drought, the WRF–LSTM chain produces lower monthly runoff and more severe SSFI values than observed, indicating the capability to represent stronger potential droughts than those in the historical record. Because LSTM does not require detailed physical characterization of the basin, the framework is transferable to data-limited or developing regions. Overall, the coupled WRF–LSTM approach provides a practical tool for exploring plausible future extreme droughts and supporting preparedness planning.

ABSTRACT. Extreme weather events have become increasingly frequent and intense under climate change, posing significant threats to public safety and disaster preparedness. Traditional weather education and hazard training often rely on static media or textual descriptions, which may limit learners’ situational understanding and risk perception. This study proposes a virtual reality (VR)-based training framework for extreme weather scenarios, aiming to enhance users’ weather hazard awareness through immersive and experiential learning. The proposed system simulates representative extreme weather events, including heavy rainfall, strong winds, and reduced visibility, allowing users to observe environmental changes and make situational judgments within a safe virtual environment. An experimental study was conducted with 35 participants, comparing the VR-based training with conventional multimedia-based instruction. Learning outcomes, risk perception, and training engagement were evaluated using pre- and post-tests as well as questionnaire measures. The results indicate that participants using the VR system demonstrated significantly higher improvements in weather hazard understanding and perceived preparedness than those in the control group. These findings suggest that immersive VR can serve as an effective supplementary tool for weather hazard education and disaster preparedness training, offering practical implications for climate adaptation and public safety education.

ABSTRACT. Numerous studies proposed disaster risk indices to support hazard assessment and risk reduction but most rely on predefined variables, with limited attention to systematic and data-driven variable selection strategies. This study addresses this gap by developing the first Typhoon Fatality Risk Index (TFRI) using variables selected through mutual information and cross-correlation and values generated from a DL-assisted, parametrically optimized hydrological model. In contrast to traditional typhoon risk assessments that primarily emphasize economic losses or flood damages, this study adopts fatality counts as the core risk indicator, thereby directly reflecting human vulnerability. The study also integrates area-specific data, Niño Sea Surface Temperature (SST) indices, and discharge values using the entropy method. To further interrogate the behavior and robustness of the selected variables, Gaussian Process Regression (GPR) is employed as an auxiliary, model-agnostic tool to examine nonlinear relationships and predictive uncertainty in fatality outcomes. Results indicate that typhoon-related fatality risk exhibits pronounced non-linear behavior, with a limited number of dominant drivers explaining a substantial fraction of outcome variability. To the author’s knowledge, this is the first study to estimate typhoon-related fatality risk by combining DL-assisted GR4J-derived discharge simulations with Niño SST indices within a unified risk index framework. The results demonstrate the strong potential of the proposed TFRI to enhance disaster risk reduction and early warning strategies, particularly in data-scarce regions where reliable impact-based risk assessments remain challenging.

ABSTRACT. Flood hazards are intensifying globally due to rapid urbanization, land-use changes, and climate variability, necessitating robust and scalable frameworks for flood risk assessment. The paper proposes a combined hydrological-hydraulic modelling framework for high-resolution flood inundation mapping in the South Pennar River basin, India, to enhance predictive capability and facilitate risk-informed decision-making. The research discusses critical issues in flood modelling, including uncertainty in input data, parameter sensitivity, and constraints in floodplain delineation. HEC-HMS is used to simulate rainfall-runoff, and HEC-RAS (1D/2D) to conduct hydraulic analysis, both combined in a GIS environment in order to characterize spatial floods. Multi-source datasets, including SRTM (30 m) and WorldDEM (12 m) Digital Elevation Models, land-use/land-cover data, soil data, and long-term rainfall records, are used. Model calibration and validation against the 2005 Villupuram flood event demonstrate good performance, with statistical indicators indicating strong agreement (Index of Agreement = 0.911, R² = 0.853, NSE = 0.57, and Pearson's coefficient = 0.923), confirming the model’s ability to reproduce observed flood dynamics. Simulation of flood inundation at various return periods (1-500 years) indicates that the extent of flooding increases significantly with return period, with the area flooded ranging from 34.63 km2 (1-year) to 344.90 km2 (500-year), which is up to 12.08 percent of the study area. The historic 2005 flood affected an area of 151.17 km² (5.29%). Projections of CMIP6 based on the EC-Earth3-Veg model are used for future climate scenarios with maximum floods of 5847 m3/s (SSP2-4.5) and 6668 m3/s (SSP5-8.5), and inundation areas of 264.98 km2 (9.28 percent) and 268.09 km2 (9.39 percent), respectively. The findings reveal a significant increase in flood risks during extreme events and climatic conditions, especially in low-lying and urbanizing areas. The research paper demonstrates that hydrological, hydraulic, and geospatial modelling are powerful tools for assessing flood risk in data-limited settings. The derived inundation maps provide essential information for infrastructure planning, early warning systems, and climate-resilient management of the basin and can be transferred to other river basins with rising hydro-climatic risks.

ABSTRACT. Climate change-induced shifts in precipitation patterns are intensifying localized rainfall events, overwhelming urban drainage systems and exacerbating flood risks. While green roofs (GR) are a recognized nature-based solution for stormwater management, their runoff reduction efficacy varies significantly depending on design specifications such as coverage fraction and soil depth. In this study, we performed numerical experiments to quantitatively analyze the stormwater retention capacity of varying green roof configurations in the Greater Seoul Metropolitan Area using historical climate data. Utilizing the Weather Research and Forecasting model coupled with a single-layer Urban Canopy Model (WRF-UCM), we simulated the hydrological responses of GRs for the period 2000–2009. The experiments were designed to evaluate the sensitivity of stormwater retention and peak discharge attenuation to two key variables: green roof fraction (50% vs. 100%) and topsoil layer thickness (10 cm vs. 25 cm). Statistical tests confirmed significant differences in performance between these configurations (p < 0.05). Our results demonstrate that increasing green roof coverage and substrate depth significantly enhances city-scale stormwater retention. We also show that this retention potential could’ve been underestimated if the antecedent soil moisture was not considered especially for extreme rainfall events. This study provides actionable insights for urban planners, suggesting that strategical optimizing soil depth in green roof guidelines is essential for enhancing urban climate resilience.

This work was supported by National Research Foundation (NRF) grants of Korea funded by the Korean government (MSIT) (RS-2022-NR072388) and Korea Environment Industry & Technology Institute (KEITI) through Climate change R&D project for new climate regime, funded by Korea Ministry of Climate, Energy and Environment (MCEE) (RS-2025-02263293).

ABSTRACT. Bangladesh faces a paradox of water abundance and water insecurity. Despite dense river networks formed by the Ganges, Brahmaputra, and Meghna rivers and high annual rainfall, access to safe drinking water remains severely constrained. Water scarcity is systemically compounded by the pronounced seasonality of precipitation, high demographic pressure, and the persistent threats of geogenic arsenic contamination and progressive saline intrusion. Climate change further intensifies these challenges through sea-level rise, stronger tropical cyclones, and rising temperatures that enhance evapotranspiration and reduce freshwater availability. By coupling physical climate-water simulations with agent-based modeling, we quantify the economic burden of water scarcity and the effectiveness of various adaptation strategies at the household level. The Water Poverty Index (WPI), the ratio of water expenditure to income, is used to spatially assess water-related economic stress. We reveal that regions with elevated salinity and arsenic contamination exhibit systematically higher WPI values, particularly in southern coastal districts such as Barisal, Khulna, and Patuakhali. Model simulations demonstrate that rainwater harvesting systems can substantially reduce household water costs and alleviate water poverty across large portions of the study area. By identifying spatial priorities and comparing adaptation options, this study highlights the effectiveness of nature-based and decentralized solutions for reducing water poverty under compound climate and water-quality risks. These findings provide a foundation for guiding targeted investments that translate climate-risk information into actionable strategies for water security and resilience in vulnerable coastal regions.

ABSTRACT. Reliable access to safe drinking water in remote and strategically sensitive areas remains a critical challenge due to limited surface water resources, infrastructure constraints and increasing climate variability. This study presents the development, deployment, and performance evaluation of a hybrid drinking water treatment and supply system that integrates rainwater harvesting and groundwater abstraction, designed as a modular, containerised treatment unit. The system was prototyped and operated at the National Water Research Institute of Malaysia (NAHRIM) as a proof-of-concept prior to upscaling for implementation in a high-profile strategic project in Sabah. The hybrid system adopts a dual-source approach to enhance supply resilience, treating rainwater and raw groundwater separately through a multi-stage treatment train consisting of sedimentation, multimedia filtration, activated carbon filtration, ultrafiltration, reverse osmosis, ultraviolet disinfection, and final disinfection. Treated water from both sources is subsequently stored and distributed for potable use. The containerised design enables rapid deployment, operational control, and ease of maintenance, making it suitable for decentralised and off-grid applications. Water quality performance was evaluated through comprehensive laboratory testing covering key physical, chemical, and microbiological parameters. Results indicate that treated water from both sources consistently complied with the Malaysian Drinking Water Quality Standards issued by the Ministry of Health. The integration of rainwater and groundwater sources also demonstrated enhanced system reliability by mitigating seasonal rainfall variability and fluctuations in groundwater quality. Overall, the findings confirm the technical feasibility and robustness of hybrid rainwater–groundwater systems as an alternative potable water solution for remote and constrained environments. The approach contributes to improved water security and climate adaptation by diversifying water sources and reducing reliance on single-source supplies. This study provides practical evidence to support the wider adoption of modular hybrid water systems in climate-vulnerable and hard-to-access regions.

ABSTRACT. The ongoing climate crisis has brought increasing attention to the potential use of groundwater resources. This study analyzes the spatio-temporal variability of groundwater monitoring data (groundwater level, EC (electrical conductivity), and temperature), collected from monitoring wells installed in fractured rock aquifers over South Korea. Using data from 234 groundwater monitoring wells, long-term trends were evaluated through non-parametric trend analyses, including the Mann–Kendall test and Sen’s slope estimator. In addition, cross-correlation analyses were conducted to assess the relationships between seasonal climate variables and groundwater responses, with statistical significance tested at the 95% confidence level. The trend analysis of groundwater levels and EC revealed increasing trends along the western and southern coastal regions, whereas inland and eastern coastal regions exhibited predominantly decreasing trends. These contrasting patterns were particularly evident when comparing inland monitoring wells with coastal monitoring wells, the latter showing substantially larger annual fluctuations. This disparity reflects the influence of sea-level rise driven by climate change, indicating an increased likelihood of seawater intrusion into coastal aquifers. Regarding climate-driven variability, cumulative rainfall was found to explain groundwater level fluctuations more effectively than individual precipitation events. Seasonal analyses showed that in winter, correlations weakened due to reduced groundwater recharge caused by low temperatures. In spring, a negative correlation between rainfall and groundwater levels was observed, likely due to low precipitation and the contribution of snowmelt from higher elevations. Summer exhibited the most unstable variability, influenced by irregular and intense rainfall combined with increased agricultural groundwater use. It also showed the longest lag times in temperature–groundwater level correlations. In fall, groundwater levels declined steadily with minimal seasonal disturbances, representing the most stable conditions. In conclusion, the long-term trends in groundwater monitoring data are strongly influenced by the geographic characteristics of monitoring wells, while seasonal variations in climate factors primarily affect groundwater level changes. These findings suggest that, as the climate crisis intensifies, future groundwater management strategies will increasingly need to incorporate seasonal response characteristics.

ABSTRACT. Impact-based early warning systems increasingly require not only reliable hydro-climatic forecasts but also information on who and what is exposed, how sensitive they are, and where adaptive capacity is weakest. This study develops a drought-focused, impact-based framework for Kenya’s Athi River Basin that integrates hydrological indicators with multi-dimensional vulnerability at county level to support more actionable early warning and response. Using daily CMIP6 precipitation projections under multiple SSP scenarios, we compute Standardized Precipitation Indices (SPI-3/6/12) to characterize historical and future meteorological droughts. Trend analysis (Mann–Kendall and Sen’s slope) is applied to identify emerging hotspots of intensifying dry conditions. These hazard metrics are combined with county-level exposure and vulnerability indicators, including population density, sectoral water demand, and adaptive capacity proxies such as water management structures and drought mitigation financing. Indicators are normalized and aggregated into a composite Drought Vulnerability Index using multi-criteria analysis, and mapped in a GIS environment. Results highlight distinct spatial patterns: downstream, counties exhibit high exposure and sensitivity to water-supply disruptions, while arid and semi-arid upstream counties show high dependence on climate-sensitive livelihoods but lower infrastructural buffering. When overlaid with SPI-based drought frequency and severity, these maps identify priority impact hotspots where even moderate hydrological deficits translate into significant socio-economic losses. We discuss how national meteorological and water authorities, county disaster units, and humanitarian partners can use these impact-based products to (i) refine drought warning thresholds, (ii) personalize alerts for different user groups, and (iii) guide anticipatory actions such as water trucking, groundwater development, and targeted social support. The approach demonstrates a scalable pathway for evolving from hazard-only drought monitoring to integrated, impact-based early warning for multi-hazard water-related risks in data-scarce regions.

ABSTRACT. Due to recent changes in rainfall patterns and increased occurrence of torrential downpours caused by climate change, damage to hydraulic structures, including reservoirs, is occurring frequently. Recent reservoir flooding caused by torrential downpours has included the collapse of Sanyang Reservoir in August 2020, and the partial loss of a levee at Wangsin Reservoir in August 2023 due to torrential downpours, forcing the evacuation of approximately 80 residents downstream of the reservoir. With reservoir-related disasters caused by torrential downpours continuing to occur, prevention of levee overflows and collapses is crucial. To prevent disasters at agricultural reservoirs, the Korea Rural Community Corporation (KRC) conducted a pilot project for disaster prevention measurement in 2023. Real-time data, including water level, flow velocity, and rainfall, are being collected for four districts. Utilizing this data, a reservoir yield prediction model has been developed and advanced research is underway to improve accuracy and ensure real-time performance for practical application. To further enhance the prediction model, the quality of training materials must be improved. In particular, to increase the reliability of model input values ​​and ensure the validity of forecast results, verification of the accuracy of rainfall gauge data and securing spatial accuracy are essential. This study integrated and analyzed rainfall, water level, flow, and reservoirs water level data collected from disaster prevention measurement pilot districts (Baekrok, Nangye, Baekryeon, and Namseonghyeon), along with neighborhood forecasts from the Korea Meteorological Administration (KMA), and high-resolution grid data. The results were used as input data for building a disaster prevention measurement-based reservoir rate prediction model. To analyze the correlation between rainfall data from disaster prevention measurement districts and KMA forecast data, neighborhood forecast grid points were identified near the four disaster prevention measurement pilot districts. Correlation analysis was conducted between the disaster prevention measurement rain gauges and each grid point within the four pilot districts, using the Pearson correlation coefficient. Based on this analysis, the automated machine learning library TPOT was applied to improve the accuracy of forecasted rainfall. Through this, we confirmed that the performance of the reservoirs water rate prediction model has improved overall compared to previous models. The method for optimizing forecast rainfall in agricultural reservoir basins developed in this study is expected to be utilized to establish a highly accurate real-time measurement-based forecasting system through improved forecast rainfall and to enhance the effectiveness of disaster response.

ABSTRACT. The impact of climate change and high urbanization is leading to increased urban flooding in coastal megacities. This paper provides a combined evaluation of the effects of climate change on flood characteristics in the Chennai basin in India, a highly vulnerable urban basin that is frequently hit by intense rainfall events. The overall aim is to measure how extreme precipitation varies and the implications of these variations for the nature of flood inundation under extreme weather conditions. Major issues that have been solved are climate model output biases, non-stationary rainfall extremes, and the simplicity of modeling urban flood processes. Redundant data on precipitation on a daily basis between CMIP6 EC-Earth3 and IMD were bias-corrected with a Relative Bias Correction Factor (RBCF), with special attention paid to an extreme rainfall event (99.99th percentile). The analysis of future projections was made under three Shared Socioeconomic Pathways (SSP 2.6, 4.5, and 8.5) between 2015 and 2100. Intensity Duration Frequency (IDF) curves imply a steadily growing intensity of rain in all the intervals of return. A two-dimensional hydrodynamic model was formulated in HEC-RAS to simulate flood inundation under current and future climate conditions with return periods of 50, 100, 150, and 200 years. Findings indicate that the area of flood inundation under SSP 2.6 and SSP 4.5 (145.52 to 163.98 km2 and 162.08 to 182.83 km2, respectively) increases by approximately 12.7% and 12.8%, respectively, with increasing return periods. The area affected by the flood is 240.05 km2 in SSP 8.5, indicating that the flood spread is significantly larger than at lower emission levels. The depth and velocity of floods also increase significantly with rainfall intensity, and deeper inundation and higher flood flow velocity were observed during high-return-period events, signifying a high risk of structural damage and urban interference. As indicated in the findings, the hazards of urban flooding are expected to intensify under future climate conditions, especially under high-emission pathways. This highlights the dire need to be more climate-resilient in urban planning, to have more robust drainage systems, and to have more adaptive flood control plans to reduce future risks.

ABSTRACT. Wonogiri Reservoir is a multifunctional reservoir located on the main stream of the upstream Bengawan Solo River. During the 44 years that the Wonogiri reservoir has been operating, there has been a significant reduction in storage capacity of 177 million m3. Based on erosion analysis, the Wonogiri catchment area have average erosion rate of 287.73 tons/ha/year. Bathymetric surveys indicate that the effective storage capacity has decreased from 388 million m³ in 1980 to 383 million m³ in 2024. Modeling results show that sedimentation rates reach 1,483,274 m³/year, with the largest contribution originating from the Keduang Sub-watershed (42%), followed by Tirtomoyo (14%), Temon (12%), and several smaller sub-watersheds. Based on the bathymetric data survey in 2024, there are a significant decrease of the total reservoir effective storage from 388 million m³ in 1980 to 383 million m³ in 2024 or around 68% from its initial capacity. Other than the flood disaster in the rainy season, the Wonogiri catchment area also facing drought problem during the dry season. This were caused by the dried of water springs in the catchment area. Therefor the implementation of infiltration well consider as an effort to reliving the water springs in the area. The purpose of constructing infiltration well as a tool for collect and absorb rainwater into the ground. It is also to store and increase the groundwater storage and also reduce the overflow of the rainwater into drains and other water bodies, which can be used in the dry season and at the same time reducing the flood disaster. Infiltration well consider as low budget and technology friendly to implement. By integrating infiltration wells in spatial planning, we can increase groundwater infiltration and reduce surface runoff, supporting urban ecosystems' sustainability. In addition, implementing infiltration wells helps maintain environmental functions, such as providing green open space that functions for water absorption, reducing the risk of flooding, preserving groundwater resources, and improving environmental quality. The focus of implementation of infiltration well is in Keduang Sub-watershed which in 2025 there has been 95 unit constructed as a collaboration with multi-stakeholder involvement, which part of more less 1,500 unit planned for the whole Wonogiri catchment area. This implementation of infiltration well were multi-years activities which involves community, companies and government collaboration Therefore infiltration well implementation not only serve as a solution to reduce flood risks but also contribute to the sustainability of water resources in the surrounding environment.

ABSTRACT. In Korea, water-related policies and systems have mainly focused on responding and adapting to the vulnerability of the water sector under the climate crisis. In contrast, policies for achieving carbon neutrality in the water sector have not been actively developed or implemented. In particular, because the water sector is not explicitly categorized in the national greenhouse gas (GHG) inventory, it is difficult to quantitatively assess current GHG emissions from the national water sector. ​This study aims to analyze GHG emissions in the water use cycle to contribute to carbon reduction in the water sector. For this purpose, indirect GHG emissions were estimated by analyzing the energy (electricity) consumed at each stage of the water use cycle. The analyzed results were then applied to both administrative districts and watershed units to compare and evaluate energy consumption across different stages of the water use cycle. By presenting the spatially and functionally disaggregated energy use of the water use cycle, the study is expected to provide useful information for planning and implementing measures in the water management sector that align with national reduction targets.

Funding: Research for this paper was carried out under the KICT Research Program (Development of IWRM-Korea Technical Convergence Platform Based on Digital New Deal) funded by the Ministry of Science and ICT.

ABSTRACT. This study addresses the Water–Energy–Food Nexus by quantifying the impacts of precipitation deficits on rice productivity and drought-affected paddy field dynamics using high-resolution remote sensing and machine learning techniques. The study focuses on Chungcheongnam-do, a major rice-producing region that experienced recurrent agricultural droughts in 2018, 2019, and 2022, accompanied by a declining trend in rice yield per unit area. Harmonized Landsat Sentinel-2 (HLS) imagery and land use maps were used to derive vegetation- and water-related indices (NDVI, LSWI, BSI, MSI, and EVI) for paddy and barren land surfaces. A machine learning-based classification framework was developed to detect drought-affected areas where paddy fields exhibited barren-like spectral characteristics due to water stress. By integrating gridded precipitation data, precipitation thresholds associated with the spatial expansion of drought-affected areas were identified, and regression analysis was conducted to quantify the relationship between affected area extent and rice production losses. Beyond food productivity impacts, the expansion of drought-affected paddy fields implies increased irrigation demand and associated energy consumption for agricultural water supply, highlighting a critical water–energy linkage under drought conditions. The proposed framework provides spatially explicit information that can support integrated decision-making for drought risk management, irrigation planning, and energy-efficient agricultural water operations. These findings demonstrate the applicability of satellite-based drought monitoring as a scientific basis for Water–Energy–Food Nexus–oriented policies under increasing climate variability.

ABSTRACT. Approximately 97% of the United Arab Emirates (UAE) consists of desert, and the country depends heavily on imports for staple crops such as wheat, rice, maize, and barley, with rice being entirely imported. To enhance food security under potential future shortages, it is necessary to develop agricultural infrastructure that enables rice cultivation in arid environments. This requires securing sufficient water resources, suitable soils, and adopting water-saving irrigation methods such as drip and simplified irrigation, along with soil structure improvement to reduce deep percolation in sandy soils. Accordingly, this study surveyed agricultural conditions, soils, groundwater, and water resources, and identified candidate areas suitable for the expansion of rice cultivation in the UAE.

ABSTRACT. Urban water distribution systems (WDS) often operate under conditions of excessive pressure caused by topographic variations and operational requirements. This excess hydraulic head is typically dissipated through pressure reducing valves, resulting in energy loss. Recovering this surplus energy represents a valuable opportunity to enhance the sustainability and energy efficiency of urban water infrastructure.

This study investigates the potential for renewable energy recovery in water distribution systems through the application of microturbines. Theoretical principles of micro-hydropower generation in pressurized pipelines are reviewed, and a systematic framework is presented to identify feasible installation locations based on hydraulic head, flow rate, and operational constraints. The framework is applied to a large-scale urban water distribution system in Busan, Korea, using a calibrated digital hydraulic model.

Candidate locations for microturbine installation are identified by analyzing spatial distributions of pressure, flow, and elevation across the network. Potential power generation is estimated considering turbine efficiency and operational limitations. The results indicate that significant recoverable energy exists at selected pressure control points, particularly in areas with large elevation differences. For the Busan case study, multiple candidate sites demonstrate feasible power outputs at the kilowatt scale, confirming the technical viability of microturbine-based energy recovery within the existing water supply infrastructure.

The analysis highlights that integrating microturbines into water distribution systems can contribute to greenhouse gas reduction, operational cost savings, and improved energy self-sufficiency of water utilities, without compromising hydraulic performance or service reliability. The proposed approach provides practical insights for utilities seeking to implement low-impact renewable energy solutions and supports the broader transition toward sustainable and resilient urban water systems.

ABSTRACT. The objective of this study is to develop a long-term daily rainfall generation framework capable of simultaneously reproducing extreme rainfall and drought characteristics, thereby enabling realistic simulation of compound disaster scenarios. Under intensifying climate nonstationarity, rainfall processes increasingly exhibit stronger extremes, prolonged dry spells, and complex transitions between wet and dry conditions. This necessitates stochastic rainfall generators that go beyond reproducing mean behavior and explicitly represent temporal persistence, tail behavior, and regime transitions. To address this need, this study proposes a neural-network-based probabilistic rainfall generator that decomposes rainfall into occurrence and intensity components. Rainfall occurrence is modeled using a nonstationary Markov chain, where daily transition structures are learned to reproduce realistic wet–dry persistence. This formulation enables the model to capture not only short-term rainfall events but also long-duration droughts, wet spell clustering, and dry–wet regime shifts, which are critical for compound hazard assessment. Conditional on rainfall occurrence, wet-day rainfall intensity is modeled using a truncated Generalized Extreme Value (GEV) distribution. Annual extreme rainfall indicators are used to define baseline distributional structure, while the neural network predicts daily deviations around these baselines. By incorporating likelihood-based training together with explicit extreme-consistency constraints, the model ensures that extreme rainfall frequency and magnitude are reproduced without sacrificing overall stability. This design allows the tail behavior of rainfall distributions to be controlled directly rather than emerging implicitly from mean-driven fitting. During future rainfall generation, the neural network predicts the parameters governing the stochastic processes, and rainfall realizations are sampled probabilistically from the resulting distributions. This approach avoids direct autoregressive dependence on previously generated rainfall, thereby reducing error accumulation and long-term drift. As a result, the framework can generate stable long-horizon simulations in which extreme rainfall events and prolonged drought periods naturally co-occur, reflecting realistic compound disaster dynamics. Overall, the proposed framework provides a robust and interpretable approach to daily rainfall generation that explicitly targets extreme rainfall, drought persistence, and their temporal interaction. The model is well suited for climate change impact assessment, hydrologic risk analysis, and studies of compound climate-driven hazards where long-term stability and extreme-event fidelity are essential.

ABSTRACT. This study proposes a novel reservoir operation framework that integrates Release Threshold Optimization (RTO) with a Hierarchical Deferred Hedging Rule (HDHR) to enhance operational flexibility while explicitly accounting for downstream ecological conditions. The proposed strategy was applied to the Gamcheon River basin in South Korea, where multiple hydrologically parallel reservoirs are operated jointly using an aggregation–decomposition approach. Reservoir inflows were generated using a calibrated hydrological model. Three variants of the proposed strategy—HDHR, Annual RTO-HDHR, and Monthly RTO-HDHR—were evaluated and compared against conventional and modified hedging rules under multiple performance criteria. Simulation results indicate that HDHR-based strategies effectively eliminated water supply deficits, while an ecologically prioritized hedging rule achieved superior ecological flow (EF) performance at the cost of substantial demand shortages. Among the proposed alternatives, Monthly RTO-HDHR demonstrated the best overall performance, achieving the highest average storage levels while consistently meeting EF targets. These findings highlight the effectiveness of combining flexible, threshold-based operating rules with ecologically informed flow targets, offering a practical decision-support approach for multi-objective reservoir management under uncertain hydrological conditions.

ABSTRACT. Recent industrial expansion has intensified the interdependence between water and energy systems, particularly in regions with water- and power-intensive facilities such as semiconductor clusters and large-scale data centers. In South Korea, developments including the Yongin Semiconductor Cluster have significantly increased both electricity demand and water consumption, amplifying systemic vulnerability under climate crisis. At the same time, climate change is altering the frequency, severity, and duration of droughts, posing growing risks to reliable water supply and, consequently, to energy production that depends on water availability. This study investigates the water–energy nexus, especially focusing on the impacts of drought conditions on energy generation. While existing water–energy nexus studies largely concentrate on accounting of water and energy consumption, they have faced limitations in examining how changes in one resource affect the other. In particular, the feedback mechanisms through which hydrological stress propagates into energy production constraints remain insufficiently explored. To address this gap, this research adopts a dynamic analytical framework that explicitly examines the interlinked responses of water supply systems and energy generation under drought scenarios. By conceptualizing water and energy systems as interacting components, the study aims to analyze how drought intensity and duration influence energy generation across different sources within a water–energy nexus framework. This approach enables a more dynamic understanding of the water–energy relationship. The study is expected to provide insights into managing water and energy resources under drought conditions. The results can support evidence-based policymaking aimed at enhancing resource efficiency and promoting integrated water–energy management strategies under growing demand and climate-driven risks.

ABSTRACT. Understanding evapotranspiration (ET) responses of dominant cropping systems to interannual climate variability is critical for agricultural water security in water-scarce regions. The North China Plain (NCP) faces severe groundwater overexploitation, largely driven by the winter wheat–summer maize (WW–SM) rotation system’s intensive irrigation demands. A two-year field experiment (Oct. 2023–Oct. 2025) was conducted in Dacaozhuang, Xingtai, Hebei Province, with local conventional management for WW (cv. Shannong No. 28) and SM (cv. Zhengdan No. 958). Using in-situ meteorological data and the FAO-56 Penman–Monteith method, reference (ET0) and actual (ETa-FAO) evapotranspiration were calculated. Both crops’ growth cycles were divided into four phenological stages to analyze stage-specific ETa-FAO and its proportional contribution. Results showed the rotation system’s total ETa-FAO ranged 867.2–983.4 mm, with significant climatic year-type differences. In the Normal year (2024) , WW consumed ~10% more water than SM; in the Wet year (2025), SM’s proportion exceeded WW by ~20%. WW ET consistently concentrated in the jointing to heading stage, while SM’s peak shifted from the great bell to grain filling (2024) to the sowing to jointing (2025), indicating strong crop water use sensitivity to climate variability. This study quantifies ET responses to climatic variability, highlighting the need to integrate climatic year types and critical water-consuming stages into irrigation and water allocation strategies. Findings provide scientific support for climate-smart irrigation, alleviating agricultural water stress and enhancing regional water security in the NCP.

ABSTRACT. South Korea often faces opposition to dam construction, which is essential for climate change response, due to local residents' concerns about land and resource loss, changes in industrial structure, and social disconnection. In addition, the massive inflow of non‑point source pollutants from the land and livestock sectors is worsening water quality, leading to intensified algal blooms. Our research team proposes a project that, centered on dams, uses water and energy to convert livestock manure, excessive pesticide application, and other intensive agriculture practices into tree cultivation, a carbon‑sequestering activity. Tree cultivation can address problems such as the declining share of agriculture in GDP and the disappearance of rural areas caused by a shrinking farming population, improve river water quality, and generate positive effects for the necessity of dam construction in response to climate change. A test-bed project will be established for the area around the Youngju Dam floodplain, involving local governments and residents in a program to advance the agricultural system surrounding the dam and improve the water quality of the Naeseongcheon River. The main initiatives include installing water‑quality‑improving filtrations, land solar powers, hydrothermal heat energy facilities, and a smart farm on the dam’s floodplain to produce water and energy; supplying seedlings of purification trees (Paulownia); utilizing river‑adjacent farmland secured through the Ministry of Environment’s land‑acquisition program as timber plantations; and enabling local residents to maintain community life and generate income through forest‑agriculture activities such as caring for the purification trees, beekeeping for honey, and employment at a timber‑processing complex. Through these measures, the state aims to reduce pollutant loads from riparian farmland management and transform water‑polluting substances into carbon‑sequestering trees, thereby achieving a reduction in non‑point source pollution.

ABSTRACT. Changes in precipitation patterns driven by climate change are leading to more frequent occurrences of drought. In addition, hydraulically controlled structures such as dams and weirs, which artificially regulate river flows, induce hydrological variability during drought conditions. Reduced river discharge during droughts affects the physical and chemical characteristics of rivers, making it difficult for such environmental changes to result in positive impacts on water quality. Drought-induced decreases in river flow negatively affect major water-use sectors, including domestic, industrial, and agricultural activities, while increased concentrations of pollutants due to reduced dilution exert adverse impacts on aquatic ecosystems. Under conditions of water scarcity caused by drought, these factors interact in a complex manner, highlighting the urgent need for real-time monitoring and proactive response strategies.

Droughts are commonly classified into meteorological, agricultural, hydrological, and socio-economic droughts. When droughts persist over long periods, they typically begin as meteorological droughts and subsequently propagate to agricultural and hydrological droughts, ultimately leading to socio-economic impacts. Climate change is projected to cause spatially and temporally heterogeneous changes in temperature and precipitation, and alterations in hydrological patterns consequently influence river water quality. Although various drought indices have been developed according to drought type and are widely used to assess drought impacts, tools specifically designed to evaluate the environmental impacts of drought remain limited.

In this study, an environmental drought (water quality) index is applied to evaluate the impacts of drought-induced water scarcity on river water quality. The index is utilized for real-time monitoring and integrated with probabilistic forecasting to enable early detection of environmental drought impacts, thereby minimizing drought-related damages under water shortage conditions. Through rapid assessment of environmental impacts and early warning capabilities, this study is expected to support the establishment of effective management and response strategies.

Acknowledgements: This work was supported by Korea Environment Industry & Technology Institute(KEITI) through Aquatic Ecosystem Conservation Research Program(or Project), funded by Korea Ministry of Environment(MOE)(2022003050007).

ABSTRACT. In Korea, the management targets for river water during dry seasons are established by taking both environmental flow and water use into account. Environmental flow is determined by calculating the required flows for water quality and aquatic ecosystems independently and then selecting the higher value among them. Under Korea’s institutional framework, because water use is intensive across the entire nation, environmental flow serves as a critical benchmark for determining the minimum streamflow that must remain in the river. Accordingly, the regulatory system mandates that water for various uses be allocated only from the residual discharge remaining after the required environmental flow has been secured However, comprehensive assessments and management strategies regarding how rivers respond and sustain their essential functions under these environmental flow conditions remain relatively insufficient. Adaptive management offers a systematic framework to address this gap, characterized by an iterative cycle of monitoring the real-world performance of policy objectives and refining those targets through a feedback-driven adjustment process. This study emphasizes the critical importance of adaptive management for river water by monitoring and evaluating both environmental flow and actual streamflow abstractions.

Funding: Research for this paper was carried out under the KICT Research Program (Development of IWRM-Korea Technical Convergence Platform Based on Digital New Deal) funded by the Ministry of Science and ICT.

ABSTRACT. Increased flood frequency and intensity under climate change have revealed the limitations of conventional, damage-centered flood management. Despite substantial public expenditures after major floods, recovery outcomes vary widely across regions, even under similar policy frameworks. This study aims to provide an operational and policy-relevant assessment of regional flood resilience by identifying component-level strengths and vulnerabilities that influence post-disaster recovery performance. A multi-criteria decision-making framework integrating the Analytic Hierarchy Process (AHP) and Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) was applied to evaluate flood resilience across 21 regions in South Korea designated as special disaster areas in 2022. Twenty indicators were organized under four key resilience components—Redundancy, Robustness, Rapidity, and Resourcefulness (4Rs). Indicator weights were derived through an expert-based AHP survey, and regional resilience levels were quantified and ranked using TOPSIS. Two additional analyses enhanced policy applicability: (i) sensitivity testing of 4R weights to assess how changes in policy emphasis affect regional rankings, and (ii) component-specific evaluations to identify internal imbalances within regional resilience structures. Results revealed significant spatial disparities in flood resilience. Rankings were particularly sensitive to the weights of Rapidity and Resourcefulness, underscoring the importance of response capacity and resource availability in effective recovery. While some regions demonstrated balanced resilience across all components, others exhibited pronounced weaknesses in specific areas, highlighting the value of disaggregated assessments for identifying targeted policy priorities. From a policy standpoint, three implications emerge. First, uniform disaster recovery measures are inadequate; component-focused interventions are needed to address regional resilience gaps. Second, the proposed AHP–TOPSIS framework provides a transparent decision-support tool for prioritizing investments under varying objectives, such as infrastructure reinforcement or rapid response enhancement. Third, the findings advocate a shift from reactive recovery to proactive resilience-based management. By strengthening weak resilience components before disasters occur, governments can mitigate future flood impacts and improve recovery efficiency. Overall, the study offers a quantitative and adaptable framework to support evidence-based flood resilience planning and policy development at both national and regional scales.

ABSTRACT. The purpose of this study is to improve the reliability of stormwater inlet capture efficiency estimation in urban flood analysis by establishing consistency between empirical design equations and hydraulic model simulations. With climate change intensifying both the frequency and magnitude of extreme rainfall events, urban areas are increasingly exposed to surface flooding, roadway overtopping, and associated secondary impacts such as traffic disruption and infrastructure damage. In highly impervious urban environments, the performance of stormwater inlets in capturing surface runoff during rainfall events plays a critical role in mitigating flood risks and reducing surface flow accumulation. In current engineering practice, inlet capture efficiencies are commonly estimated using empirical equations provided in the HEC-22 manual. These equations are widely applied due to their simplicity; however, they are often used without systematic consideration of their consistency with dynamic hydraulic models employed in urban flood analysis. As a result, discrepancies may arise when inlet performance is represented within distributed hydrologic–hydraulic models such as the Storm Water Management Model (SWMM), potentially introducing uncertainty into flood simulation results and drainage system assessments. To address this issue, this study examines the relationship between experimentally measured inlet capture efficiencies obtained under realistic roadway flow conditions and capture efficiencies simulated using HEC-22 equations in EPA-SWMM 5.2. Experimental data reported in previous studies were digitized and used as reference observations to ensure a consistent basis for comparison. Based on the observed relationship between experimental results and SWMM simulations, the existing HEC-22 inlet capture efficiency equations were reformulated to achieve improved consistency with SWMM-based hydraulic modeling. Using the HEC-22 inlet formulation, a generalized set of inlet capture efficiency regression equations was developed to enable application across a wide range of inlet sizes and inflow conditions, rather than being limited to specific grate dimensions. The generalized equations allow inlet capture efficiency to be flexibly estimated for arbitrary inlet configurations within SWMM-based urban drainage models. The proposed framework provides a coherent approach for representing inlet capture efficiency within hydraulic modeling environments, supporting more reliable and flexible simulation of stormwater inlet behavior. By enhancing the consistency between empirical inlet design equations and hydraulic model applications, this study contributes to improved urban flood analysis and more robust stormwater drainage system design.

ABSTRACT. Effective flood management requires coordination across fragmented governance clusters, yet the institutional interdependencies connecting these clusters often remain hidden within complicated, multi-layered policy documents. This study develops an integrated analytical framework to identify two distinct types of network vulnerabilities: weak ties—critical existing connections bridging otherwise disconnected clusters—and structural holes—absent relationships whose creation would most effectively improve system integration. We extracted institutional relationships from Korean water governance documents using a rule-based text analysis approach and constructed a directed network representing actors and infrastructure components. Network analysis methods were applied to detect governance clusters and quantify both existing bridges between clusters and potential new connections that would reduce network fragmentation. Our findings reveal complementary vulnerability patterns. Weak ties in Korea's governance system function as critical linkages through central coordinating authorities, connecting national policy-making bodies with local implementation units. This concentration creates critical dependency on few coordination channels. Structural hole analysis uncovered different leverage points: emergency response actors, despite peripheral formal positions, occupy strategic locations where new institutional linkages would most effectively enhance integration across governance domains. The distinction between weak ties and structural holes proves essential for intervention design: existing weak connections require strengthening through resource allocation and protocol clarification, while structural holes demand institutional transformation to create entirely new coordination pathways. This dual diagnostic approach provides a transferable framework for enhancing flood resilience across diverse water governance contexts.

Acknowledgement This work was supported by the National Research Foundation of Korea(NRF) grant funded by the Ministry of Science and Technology (RS-2024-00356786).

ABSTRACT. Agricultural reservoirs have been designed and operated primarily for irrigation water supply and are typically managed to maintain storage levels close to full capacity. However, the need to strengthen flood response capabilities has recently increased due to the growing frequency of localized and extreme rainfall events caused by climate change, as well as the aging of reservoir infrastructure. Consequently, agricultural reservoirs are now transitioning into multi-functional water resource systems that must consider not only water supply but also disaster prevention through water level regulation during the flood season. The objective of this study is to propose a rational water level management strategy that ensures the operational safety of reservoirs during the flood season while minimizing impacts on agricultural water supply. To this end, a flood-season management level estimation method integrating both water supply (utilization) and flood control perspectives based on the water balance concept is developed. From the water supply perspective, key factors—including inflow, outflow, watershed ratio, normal storage rate, and minimum required storage—are incorporated to reflect reservoir water balance characteristics. From the flood control perspective, a simplified equation is developed that accounts for flood-season inflow characteristics and reservoir discharge capacity to derive management levels for flood mitigation. Furthermore, the management levels derived from the water supply and flood control perspectives are combined and applied to the early, middle, and late stages of the flood season to establish stage-specific water level criteria that simultaneously consider water supply reliability and flood response capability. When applied to 1,200 agricultural reservoirs nationwide, the proposed method indicated that the average management levels should be maintained at approximately 88% of total storage capacity in the early stage, 70% in the middle stage, and 78% in the late stage of the flood season. These results quantitatively demonstrate the necessity of stage-based water level regulation in reservoir operation. The findings of this study provide scientific and rational decision-making information for setting flood-season management levels of agricultural reservoirs and are expected to contribute to more efficient operation and enhanced safety in reservoir disaster management.

This work was supported by Korea Institute of Planning and Evaluation for Technology in Food, Agriculture and Forestry(IPET) through Intelligent Agricultural Infra Management for Climate Change Development Program, funded by Ministry of Agriculture, Food and Rural Affairs(MAFRA)(RS-2025-02215604).

ABSTRACT. The Saemangeum reclamation project, initiated in 1991 to establish large-scale farmland, has undergone several revisions. In 2008, the plan was adjusted to 9,430 ha (30% of the total area), and in 2021 further modified to shift from paddy-focused to dry-field farming, with completion targeted for 2030. Persistent deterioration of water quality in Saemangeum Lake led to abandoning freshwater conversion and adopting seawater inflow, raising concerns about its impacts on soil and water resources. This study investigates changes in salinity and water levels in relation to soil salinity, groundwater, and drainage conditions, aiming to propose strategies for sustainable agriculture. Monitoring and spatial analysis revealed that since 2020, seawater inflow increased lake conductivity by about 4.0 dS/m, with reservoir salinity also rising. Farmland adjacent to the lake showed slight increases in soil salinity, while reservoirs acted as buffers. After reclamation, soil salinity generally declined below 2.0 dS/m within four years, stabilizing crop cultivation. However, about 5% of the area-particularly near the lake and in silty loam zones-remains vulnerable to salt damage, requiring localized desalination and continued monitoring. Drainage analysis indicated poor conditions in some plots, with water persisting for more than seven days after rainfall. Zone-specific surveys revealed heterogeneous performance: Zones 5 and 7-2 were rated A (good), whereas Zones 2, 3, and 6-2 were rated D~E, indicating severe drainage problems linked to soil properties, topography, and management practices. Groundwater monitoring showed that heavy rainfall raised groundwater close to the surface, with conductivity remaining high. A significant correlation between groundwater and soil conductivity confirmed groundwater salinity as a key driver of soil salinity. Modeling demonstrated that reservoir water-level control effectively lowered groundwater levels, and freshwater conditions reduced salinity overall. In contrast, seawater inflow increased salinity and re-salinization risks, with long-term accumulation possible in upper groundwater layers. Based on these findings, strategies are proposed: cultivating salt-tolerant crops, operating reservoirs under freshwater conditions, installing subsurface drains, and applying soil amendments for localized desalination. Drainage improvement through land leveling, channel installation, and deep tillage should be combined with improved management practices. Groundwater management measures include crop selection and adjusted cultivation schedules, supported by zone-specific plans. This study provides scientific evidence on the agricultural impacts of seawater inflow management in Saemangeum Lake, offering practical reference data for policy decisions and contributing to resilient agricultural models under climate and water resource uncertainties.

ABSTRACT. Among the various natural disasters affecting Korea, floods and droughts have caused the most severe damage. In the future, the scale and frequency of such disasters are expected to increase further due to the impacts of climate change. The development of systems and technologies for flood and drought response in Korea can be said to have begun with the establishment of the Han River Flood Control Office in 1974. Since then, flood control offices have also been established in other river basins, including the Nakdong River, and continuous efforts have been made to advance flood control and water management. In recent years, the role of flood control offices has become increasingly important across a wide range of areas, including flood control, drought management, river water use management, and international cooperation. At the same time, active efforts are being made to apply advanced technologies such as artificial intelligence and digital twins to water management in order to improve the accuracy of flood and drought forecasting and to strengthen proactive response capabilities. Accordingly, this study presents the evolution of flood forecasting operations in Korea, centered on the nation’s flood control offices, and examines how water management practices have developed over time. The current status of water management tasks carried out by flood control offices is organized by sector, and future environmental changes and development directions are proposed. Through this, the study provides foundational information aimed at protecting the public from water-related disasters and enhancing national water welfare.

ABSTRACT. Urban flood management is increasingly challenged by the limited capacity and adaptability of conventional gray infrastructure under rapid urbanization and climate change. In this context, hybrid nature-based solutions (NbS), which integrate green measures with gray and, in some cases, blue infrastructure elements, gain attention as integrated approaches for urban flood mitigation. By coupling engineered flood control with ecosystem-based functions, these systems aim to enhance hydraulic reliability while improving adaptability to changing hydrological conditions. A systematic review of existing literature is conducted to examine how NbS-based hybrid flood management systems are evaluated and discussed in urban contexts. The review indicates that hybrid systems outperform single-measure approaches; however, practical constraints affect flood risk reduction and system performance across conditions and time scales. Empirical evidence on system behavior during extreme rainfall events and over long operational periods remains limited, raising concerns about the reliability of integrated urban flood solutions. While many studies emphasize conceptual advantages or short-term outcomes, systematic evaluation of extreme-event response, long-term performance degradation, maintenance requirements, and failure mechanisms requires further consideration. The results further show that critical issues such as optimal gray–green ratios and cumulative impacts under repeated flooding remain insufficiently explored. These findings inform future decision-making by clarifying key uncertainties that affect the long-term reliability and resilience of NbS-based hybrid urban flood management systems.

Ackknowledgement:This work was supported by Korea Environment Industry & Technology Institute(KEITI) through Climate Resilient R&D Project for Water-Related Disaster Management, funded by Korea Ministry of Climate, Energy and Environment(RS-2023-00218973)

ABSTRACT. Agricultural reservoirs play a critical role in supplying irrigation water and mitigating drought impacts. However, in many small- and medium-sized agricultural reservoirs in Korea, reservoir inflow and outflow are not directly measured due to practical constraints. This lack of direct observations limits the availability of fundamental hydrological data required for reliable reservoir operation, and uncertainty in inflow information can undermine operational decision-making. In contrast, reservoir water levels are relatively easy to monitor using water level gauges, and changes in storage can be estimated through established water level?storage relationships. Nevertheless, when water level?based water balance equations are applied at an hourly time scale, the estimated inflow time series often exhibit excessive fluctuations and physically unrealistic negative values. These issues mainly arise from measurement errors in water level observations and uncertainties in the water level?storage relationship, and they become more pronounced as temporal resolution increases. To address these limitations, this study develops a reservoir inflow estimation framework that integrates hourly water level observations with Kalman filter?based state estimation techniques. In particular, the Kalman Filter (KF) and the Ensemble Kalman Filter (EnKF) are employed as data assimilation approaches that can probabilistically account for observation errors and uncertainties in state variables. The proposed framework incorporates the assumption that reservoir inflow varies smoothly over short time intervals into the state equation, while hourly water level observations are assimilated through the observation equation. This structure enables the simultaneous estimation of reservoir inflow and storage while suppressing nonphysical inflow variability. A state-space model is constructed by defining reservoir inflow and storage as state variables, and the water level?storage relationship is applied as the observation model. Hourly inflow estimates are sequentially updated using the KF or EnKF. The performance of the proposed method is evaluated through comparisons with conventional water level?based inflow estimation approaches, focusing on reductions in inflow fluctuations and improvements in physical consistency. In addition, observed inflow data available at selected reservoirs are used to further assess the accuracy and validity of the Kalman filter?based estimates. The results demonstrate that the proposed framework effectively reduces nonphysical fluctuations in hourly inflow estimates and improves the stability and reliability of inflow time series derived solely from water level observations. This study provides a practical solution for inflow estimation in agricultural reservoirs with limited hydrological measurements and offers a useful foundation for real-time reservoir operation and future integration with hydrological and reservoir operation models.

ABSTRACT. In paddy fields, both rainfall and irrigation water supplied from reservoirs interactively contribute to the hydrological cycle. Quantitative decomposition of return flow according to its source provides a critical basis for diagnosing irrigation operation efficiency and evaluating the rationality of agricultural water management. However, conventional return flow analyses largely rely on aggregated volumes or simple return flow ratios, which limits process-based interpretation of internal hydrological behavior and operational performance.

In this study, a source-based hydrologic component decomposition framework is developed using the Dynamic Water Resources Assessment Tool (DWAT), a physically based semi-distributed hydrologic model. Major hydrological components in paddy fields—including surface runoff, interflow, baseflow, infiltration, evapotranspiration, and water storage—are simulated and explicitly separated into rainfall-origin and irrigation-origin contributions. This approach enables quantitative assessment of source-specific hydrologic responses to irrigation operations.

The proposed framework allows spatio-temporal analysis of individual hydrologic components and provides a distinctive capability to track source-specific water storage dynamics within soil and groundwater layers on a monthly basis. Such storage-oriented analysis offers deeper insight into water retention, loss pathways, and return flow generation processes that are not adequately captured in conventional ratio-based assessments.

Through this decomposition, irrigation operations characterized by rapid conversion of supplied water into surface runoff can be diagnosed as hydrologically inefficient, whereas sustained storage and utilization of rainfall-origin water indicate potential for optimized supply strategies. These diagnostic outcomes enable qualitative, process-based evaluation of irrigation performance that cannot be inferred from return flow metrics alone.

The proposed methodology can support the development of improved reservoir operation strategies and irrigation water allocation policies. By enhancing understanding of source-specific hydrologic pathways, this study contributes to sustainable agricultural water management and strengthens adaptive operational capacity under increasing hydrological uncertainty driven by climate variability.

This work was supported by Korea Environment Industry & Technology Institute(KEITI) through Water Management Program for Drought Project, funded by Korea Ministry of Climate, Energy and Environment(MCEE).(2022003610002)

ABSTRACT. This paper systematically reviews the background, revision history, and applicability of China’s evaluation standard for green small hydropower stations. It focuses on analyzing a comprehensive quantitative evaluation index system structured around "three-tier, four-dimensional" framework, comprising 4 categories, 14 elements, and 21 indicators, aiming at achieving eco-environmental friendliness, social harmony, standardized management, and cost-effectiveness. The paper highlights its seven core requirements and six technical innovations. Through practical verification across over 1,000 power stations, this standard has demonstrated extensive adaptability and excellent practicality, yielding significant outcomes in promoting river ecological restoration, advancing workplace safety standardization, and supporting rural revitalization. This framework offers a valuable insights for the sustainable development of small hydropower stations globally. Notably, only 2.6% of power stations have undergone green small hydropower station evaluation, underscoring the urgent need for enhanced technical promotion and policy incentives, ISO-aligned cross-border certification recognition, and transnational certification mechanisms.

ABSTRACT. The Cheonggyecheon Stream, a restored urban waterway in Seoul, experiences recurrent water quality degradation during rainfall events due to inflows of combined sewer overflows (CSOs). While sluice gates along the stream are primarily operated to prevent inland flooding, the current operation scheme allows highly polluted first-flush runoff to be discharged directly into the stream. This study investigates a sluice gate operation strategy that strategically delays gate opening to utilize the storage capacity of the sewer system, aiming to improve water quality protection while maintaining hydraulic safety. In this study, XP-SWMM (Storm Water Management Model) simulations were conducted to analyze the characteristics of the Cheonggyecheon catchment. Using design and as-built information of the covered structures, a detailed hydraulic model consisting of approximately 10,000 nodes and links was developed. Dry-weather sanitary flow conditions were established based on real-time water flow data obtained from the monitoring system. Operation scenarios were defined according to different overflow weir heights within the covered structures, and the effects were evaluated in terms of reductions in gate opening frequency and changes in sewer water levels. Simulation results indicate that under minor rainfall conditions, sections with limited hydraulic capacity can be operated with delayed gate opening while maintaining water levels below critical thresholds. In these upstream reaches, increasing allowable overflow heights enables effective retention of CSOs during the initial stage of rainfall without inducing upstream flooding. However, under higher rainfall intensities, water levels rise rapidly, necessitating timely gate opening to ensure hydraulic safety. Overall, this study provides a scientific basis for optimizing gate operations, suggesting that data-driven, delayed opening protocols can enhance the ecological resilience of restored urban streams without compromising urban flood safety.

ABSTRACT. Recent increases in the frequency and intensity of localized extreme rainfall events, driven by climate change, have exacerbated urban flood risks. Concurrently, rapid urbanization has expanded impervious surfaces in cities, frequently exceeding the design capacity of existing drainage systems and leading to recurrent flood damages. These impacts manifest as road inundation, traffic disruption, and degradation of urban system functionality. Conventional urban flood assessments, which rely on single indicators such as maximum inundation depth or flooded area, fail to capture the temporal degradation of urban system performance during flooding and the subsequent recovery process. Consequently, there is a growing need for urban flood resilience assessments that explicitly consider system performance dynamics before, during, and after flood events. In this study, system performance were defined separately for road networks and residential areas, and urban flood resilience was quantified by distinguishing resistance and recovery components to reflect temporal variations in system performance. In addition, indicators such as performance loss, residual performance, impact duration, and recovery duration were used to evaluate the vulnerability and recovery characteristics of urban systems. To do so, a one-dimensional (1D) runoff model was developed using the Storm Water Management Model (SWMM) of the United States Environmental Protection Agency, and a two-dimensional (2D) inundation model was constructed using LISFLOOD-FP. The coupled 1D–2D modeling framework was employed to derive system performance curves based on inundation depth and flow velocity, enabling a quantitative assessment of urban flood resilience. The results of this study provide a scientific basis for resilience-based urban flood risk assessment and offer practical insights to support flood management strategies and urban planning under increasing climate extremes.

ABSTRACT. Water management is shaped by ongoing interactions between hydrological processes and human decision-making. Both the ecosystem services (ESS) framework and sociohydrology seek to conceptualize these interactions by linking biophysical processes with societal values and institutional responses to support sustainable natural resources management. Although the two approaches have largely developed in parallel, we show that they share a common structure: hydrological processes generate services, these services shape social priorities and management decisions, and these decisions feed back into the water system. This shared causal architecture suggests that ecosystem services are integral, endogenous components of coupled human–water systems rather than external outcomes. Building on this insight, we develop a conceptual sociohydrological framework that explicitly embeds hydrological ecosystem services within system feedbacks to strengthen sociohydrological analysis and its relevance for integrated water resources management (IWRM). In this framework, ecosystem services are reframed as dynamic system properties link hydrological conditions with institutional decision-making, rather than being treated as static or purely economic indicators. The proposed framework consists of four interacting subsystems; hydrology, institution, technology, and socioeconomy, and provides an operational structure for translating hydrological states into socially relevant information. Outputs from the hydrology subsystem, such as runoff, soil moisture, and moisture deficit, are translated into indicators of provisioning, regulating, supporting, and cultural ecosystem services. These indicators are further interpreted as system-level resilience capacities, including resistance, recovery, and reliability, using a sociohydrological resilience canvas. Preliminary application of the framework demonstrates how hydrological variability can be systematically translated into decision-relevant resilience information that connects physical water dynamics with socio-institutional responses. By operationalizing ecosystem services as dynamic components of sociohydrological systems, the framework provides a conceptual foundation for integrating ecosystem services into sociohydrology and supports more adaptive and resilient approaches to IWRM.

ABSTRACT. Quantitative assessment of long-term hydrological components (precipitation, actual evapotranspiration, runoff and groundwater recharge) is essential for understanding structural shifts in the water cycle under climate and land use change and for establishing integrated water resources management strategies. This study aims to simulate long-term hydrological components across South Korean river basins using the SWAT-K and to analyze their spatial distribution and temporal variability. SWAT-K were constructed, calibrated, and validated using long-term meteorological and streamflow observations. Annual and monthly water balance components (precipitation, evapotranspiration, runoff and groundwater recharge) were quantified, and variations in hydrological component ratios were examined. The results provide a comprehensive diagnosis of the vulnerability and resilience of South Korean river basins and offer a scientific basis for climate-adaptive water resource planning and the development of the IWRM-K (Integrated Water Resources Management–Korea) platform.

Acknowledgements: Research for this paper was carried out under the KICT Research Program (project no. 20260156-001, Development of IWRM Korea Technical Convergence Platform Based on Digital New Deal) funded by the Ministry of Science and ICT.

ABSTRACT. Water infrastructure such as dams, floodgates, and pumping stations requires accurate and efficient detection and monitoring to ensure structural integrity and safe operation. As these structures age and deteriorate, continuous monitoring and early warning systems become essential. Traditional manual inspections (e.g., visual crack surveys by experts) are labor-intensive, slow, and often impractical. Thus, automated sensor-based solutions offering 24/7 monitoring are increasingly adopted, as they can detect subtle changes that human observers might miss. We present a unified framework to evaluate the performance of both physical and AI-based virtual sensors. Traditional physical sensors (e.g., triaxial accelerometers, gyroscopes, magnetometers) are assessed according to international standards, with metrics such as sensitivity, frequency response, linearity, repeatability, and noise. Virtual sensors using AI to detect cracks from CCTV footage are evaluated via metrics like accuracy, recall, and false positive/negative rates by comparing detection results with ground-truth data from RTSP video streams. These two approaches are complementary, and the unified evaluation framework highlights their combined value for more reliable water infrastructure monitoring.

Acknowledgement : This work was supported by Korea Environmental Industry & Technology Institute(KEITI) Through Development of an Early Disaster Detection and Response Decision Support System for Water Resources Facilities Based on Physical and Virtual Sensor Networks Project, funded by Ministry of Climate, Energy, Environment(MCEE).(RS-2024-00331809)

ABSTRACT. Floating debris generated in Korean rivers has a significant impact during flood events. When accumulated at transverse hydraulic structures such as bridges, floating debris blocks sunlight, thereby degrading riverine and marine ecosystems and deteriorating landscape aesthetics. In addition, floating debris tends to accumulate easily around hydraulic structures, causing adverse hydraulic effects such as increased flood water elevations and enhanced local scour. As a result, floating debris negatively affects the aquatic environment and increases the structural risk and instability of water resource facilities. In this study, an Artificial Neural Network (ANN)–based prediction model was developed to estimate the annual amount of floating debris generated in Korean rivers. Thirteen multipurpose dams were selected as study areas and classified according to major river basins in Korea. Various influencing factors, including hydrological, geometric, and water quality variables, were applied to the model by basin. The prediction results demonstrated that the proposed model achieved more accurate and reliable predictions of floating debris generation compared to previous studies.

ABSTRACT. Facility operators and regulators increasingly require facility-level demand accounts that can be reviewed, updated, and compared across assets. Yet in many jurisdictions, demand estimation is still anchored to fixed unit-demand tables, which can lag behind demographic transitions and shifts in industrial activity. At the same time, water-balance assessments often concentrate on deficit detection, while guidance for consistent, auditable demand evaluation is scattered across documents and practices. This study proposes a standardized framework that organizes the institutional requirements, data linkages, and reporting rules needed for routine facility re-evaluation and for future extensions toward surplus accounting. The framework consists of three parts. First, domestic legal requirements and widely used international planning standards are examined to define essential procedural steps and minimum documentation for facility re-evaluation. Second, a process architecture specifies common indicators, quality-control checkpoints, and reporting templates so that analyses remain comparable across facilities and across years. Third, a relational facility–service-area database captures how supply facilities serve domestic, industrial, and agricultural demand zones, including shared supplies and hierarchical distribution routes represented through many-to-many linkages. Baseline demand modules follow unit-demand formulations consistent with prevailing guidance. Domestic demand is computed from demographic and service-coverage attributes; industrial demand is derived from selectable unit-demand options linked to spatial development or economic activity and screened against historical withdrawals; and agricultural demand is estimated from cultivated area using standardized unit water requirements. Inputs are assembled from public statistical and hydrometeorological data services and curated archives, then processed with plausibility checks, missing-data handling, and versioned metadata to preserve traceability. Supply-side descriptors—facility specifications, storage and release records, streamflow observations, and environmental-flow obligations—are cataloged to support consistent accounting of operational constraints. The framework delivers standardized documentation, linkage tables, geospatial allocation layers, baseline datasets, and reusable computational modules, enabling transparent review, inter-facility comparison, and reproducible updates for water-resources planning.

ABSTRACT. Identification of the natural flow regime is fundamental to assessing water availability, sustaining aquatic ecosystems, and evaluating the impacts of hydro-infrastructure. This study estimates naturalized flow duration curves (FDCs) for the Gamcheon Stream, South Korea, which has been extensively altered by water uses and discharges, including stream water withdrawals, groundwater withdrawals, treated wastewater discharges, and releases from a multipurpose dam. A semi-distributed hydrological model, SWAT (Soil and Water Assessment Tool), was implemented and calibrated using observed streamflow data under anthropogenic activities, after which naturalized streamflows were simulated by excluding water uses and operational releases. FDCs and low-flow statistics were then derived at multiple points along the stream. The results show that FDCs based on simulated naturalized streamflows differ from those estimated using simple drainage-area ratio methods, and that groundwater abstraction exerts a dominant influence on low-flow characteristics.

Funding: This work was carried out under the KICT Research Program (Development of IWRM Korea Technical Convergence Platform Based on Digital New Deal, Grant number 20260156-001) funded by the Ministry of Science and ICT.

ABSTRACT. As urban areas face increasing flood risks due to climate change and rapid development, assessing the resilience of drainage systems becomes essential. This study introduces a method to evaluate the flood resilience of urban drainage networks (UDNs) using structural data and flood predictions. A deep learning model combines Graph Neural Networks (GNNs) and Transformer encoders to predict water depth at each network node. The GNN captures spatial dependencies within the drainage system, while the Transformer models rainfall-runoff dynamics over time under various storm conditions. Flooded areas are identified by setting a depth threshold, creating dynamic flood maps for the system. Resilience is evaluated at the pipe level using an adapted Simple Urban Flood Resilience Index (SUFRI), focusing on three key indicators: pipe depth (maximum flood depth a pipe can withstand), pipe recovery time (time taken to return to normal after flooding), and pipe frequency (how often flooding occurs). These indicators are used to calculate resilience scores for each pipe, which are then weighted according to their importance in the network, considering factors like flow capacity, location, and redundancy. The overall system resilience score is derived by aggregating the weighted scores of all pipes. This methodology is applied to a real-world drainage system, evaluating resilience under multiple storm scenarios. The analysis identifies areas with low resilience and pipes that require upgrades. The framework suggests improvements, such as increasing pipe capacity, enhancing redundancy, and optimizing recovery infrastructure, to mitigate future flood risks. By integrating advanced flood prediction models with pipe-level resilience metrics, this approach provides a scalable and data-efficient tool for managing flood risk in urban areas with limited data. Acknowledgement This work was supported by the National Research Foundation of Korea(NRF) grant funded by the Ministry of Science and Technology (RS-2024-00356786) and Korea Environmental Industry & Technology Institute grant funded by the Ministry of Environment (RS-2023-00218973).

ABSTRACT. The core challenge in urban sustainable development lies in quantitatively analyzing the complex coupling relationships among multidimensional subsystems, namely socio-economy, water resources, energy, and ecological environment (W-E-E-S). Although existing research has made significant progress in sustainability assessment, most frameworks still focus on the traditional three-dimensional perspective of society-economy-environment or are limited to binary correlation analysis of resource and environmental elements. They often fail to place water and energy subsystems on an equal footing with socio-economic and ecological environments, thus making it difficult to systematically reveal the synergistic and antagonistic mechanisms within the complex giant system of megacities. Therefore, this study constructs an urban sustainable development assessment framework integrating the W-E-E-S four-dimensional perspective. This framework first establishes a multi-system comprehensive evaluation indicator system applicable to the urban scale. It then introduces a coupling coordination degree model to quantitatively characterize the synergistic evolution trajectory and coordinated development level of interactions among the subsystems. Furthermore, an obstacle degree model is applied to diagnose the key factors constraining the system's coordinated development. An empirical study using Wuhan as a typical case aims to reveal the internal driving mechanisms and evolution patterns of its sustainable development level. The findings are expected to yield a more profound understanding of the interactive dynamics within such complex urban systems. Ultimately, this research seeks to establish a robust scientific basis, offering evidence-based insights for formulating targeted management strategies. This is crucial for optimizing the high-quality development pathways not only for Wuhan but also for other similar megacities characterized by comparable resource and environmental endowment profiles.

ABSTRACT. This study aims to support advanced dam operation by proposing a dam inflow prediction method using the machine learning technique Elastic Net (EN). In recent years, in Japan, climate change has intensified extreme events such as heavy rainfall and droughts, leading to increasingly diverse and sophisticated demands on dam operation. Particularly in summer, in addition to water-use constraints caused by low rainfall, the difficulty of flood control has increased due to localized heavy rainfall and the frequent occurrence of linear precipitation bands. Under these conditions, highly accurate inflow prediction is essential for advanced operational decisions such as reducing ineffective releases and conducting early pre-releases. Furthermore, toward the realization of a carbon-neutral society, expectations are growing for hybrid dam operation that integratively manages low-flow conditions for water use and hydropower generation in addition to flood control. In this study, Takadomari Dam, a water-use dam located in Hokkaido, Japan, was selected as the target, and two types of inflow prediction models based on EN were developed to predict dam inflow. The first method uses a single model trained with all available information from the summer flood season of a given year (full-period training model). The second method constructs models specialized for high-flow and low-flow conditions and switches between them using the abundant flow rate, an index used in dam operation, as a threshold (high-flow and low-flow switching model). The prediction periods were the summer flood season (July–October) of 2014, when the largest historical peak inflow was observed, and the summer flood season (July–October) of 2025, which showed relatively low-flow conditions. The prediction results indicate that, for continuous inflow prediction throughout the summer flood season, both EN-based methods were able to predict flow regime variations well. In addition, prediction using the full-period training model (2023) tended to produce larger values than the observed inflow overall, representing a flood-control-oriented (conservative) tendency. In contrast, prediction using the high-flow and low-flow switching model showed reduced overestimation and improved prediction waveforms, indicating a water-use-oriented safety tendency.

ABSTRACT. Traditional water allocation is often treated as a top-down optimization problem, frequently overlooking the inherent spatial advantages of upstream users. Stakeholders with upstream intake facilities possess a natural priority, creating benefit-cost asymmetries across the basin. This study proposes a framework to quantify these disparities among upstream, midstream, and downstream users and introduces a stage-based intake restriction strategy to promote water justice during droughts. The model is applied to the Yeongsan River Basin in South Korea, where community-based research demonstrates how collaborative intake policies can enhance both net profit and equity. A key contribution of this research is the integration of multi-objective optimization with a game-theoretical approach to evaluate the feasibility of Pareto-optimal solutions. Specifically, the analysis examines whether the marginal profits gained by downstream users are sufficient to compensate upstream and midstream stakeholders for their cooperation. To ensure long-term applicability, the framework is extended to projected scenarios of intensified climatic and socio-economic droughts. This assessment confirms the effectiveness of the proposed solutions under future stress conditions. By identifying policies that align individual incentives with collective goals, this study provides a practical template for achieving efficiency, equity, and sustainability in agricultural water governance.

ABSTRACT. The Sihwa Reclamation Project was initiated as a national development program aimed at securing agricultural land and water resources, while promoting industrial, urban, and recreational development along Korea’s western coast. However, due to severe water quality degradation in the planned freshwater lake, the project was revised to convert it into a seawater lake (Sihwa Lake, 3,620 ha). The reclaimed land was primarily allocated for industrial and residential use (9,734 ha), with a portion designated for agricultural land (3,636 ha) and a secondary freshwater reservoir, Tando Lake (760 ha). Tando Lake, with its limited watershed area, was considered incapable of achieving desalination through natural inflow alone. To address this, a plan was established in 2001 to supply desalination water from Hwaseong Lake, created through the adjacent Hwaong Reclamation Project. However, due to water quality issues in Hwaseong Lake, the supply was delayed. Despite this, Tando Lake, completed in 2010, exhibited a gradual decrease in salinity even without external desalination water, maintaining levels below the target threshold of 1,280 ppm by 2023. This study evaluates whether Tando Lake can maintain the target salinity required for agricultural water supply without external input. To estimate the required desalination water, the DESAL model—based on the Minami salt balance equation—was applied in 2010 to simulate salinity before and after supply. Results indicated that without desalination water, average salinity would remain at 2,058 ppm, exceeding the target. In contrast, with desalination water, average salinity stabilized at 819 ppm, consistently meeting the target during the irrigation period (May–September). Long-term monitoring from 2001 to 2025 showed a steady decline until 2013, followed by stabilization around 2,000 ppm. Between August 2022 and December 2024, salinity remained below the target but rose slightly in 2025. Correlation analysis revealed a weak relationship between annual rainfall and salinity (r² = 0.52), but a strong correlation with two-year average rainfall (r² = 0.78). The required two-year average rainfall to meet the target salinity was estimated at 1,566 mm, a threshold exceeded only 9% of the time over 45 years. To ensure stable agricultural water supply, alternatives were evaluated, including desalination water from Hwaseong Lake or treated effluent (80,000 m³/day) from a nearby wastewater treatment plant. Both options face economic and environmental challenges. As a feasible alternative, this study proposes operating Tando Lake as a brackish reservoir for emergency use, while supplying only the required agricultural water through additional treatment of effluent.

ABSTRACT. Flood-risk governance requires integrated water management because disaster impacts propagate across administrative boundaries and infrastructure responsibilities. However, assessments often isolate nationally managed rivers, local rivers managed by regional governments (metropolitan cities and provinces), or urban pluvial flooding (urban drainage flooding), and emphasize asset-level direct damages, potentially understating system-wide losses and cross-jurisdictional dependencies. We quantify comprehensive economic flood losses for the metropolitan region (Seoul, Incheon, and Gyeonggi) of South Korea by integrating object-based direct-damage estimates and interregional indirect losses to study multi-level flood risk governance. Direct damages for buildings, population, vehicles, and agricultural parcels are computed from inundation depth using depth–damage and unit-loss functions across multiple return periods and aggregated as expected annual damage (EAD). For flood scenarios, direct losses are translated into region–sector shocks and outward interregional spillover losses are simulated using an interregional input–output (IRIO) model, tracking propagation from Seoul, Incheon, and Gyeonggi to other regions. We compare five multi-component flood scenario configurations: three single-component cases, a combined fluvial case (national & local), and a combined fluvial–pluvial case (national & local & urban). Losses vary markedly by configuration and increase nonlinearly under combined scenarios; in the fully combined case, indirect losses account for 35.44%, 39.60%, and 38.85% of total losses in Seoul, Incheon, and Gyeonggi, respectively, revealing dependencies not visible in direct damages alone. To address post-flood landslide risk under extreme rainfall, we compare direct losses only under three event scenarios—flood-only, landslide-only, and landslide-after-flood (post-flood landslides)—and show that post-flood landslides can compound local direct losses beyond flood impacts. These findings underscore the need for coordinated planning across river management, urban drainage, and slope/land-management responsibilities when prioritizing integrated flood-risk reduction in the Seoul metropolitan region, Korea.

ABSTRACT. Amid intensifying droughts driven by climate change, groundwater dams serve as vital infrastructure for water security. However, the advanced age of South Korea’s some facilities necessitates urgent, systematic maintenance. Because the underground cut-off barriers the systems' primary components are not visible, their condition is difficult to assess. Thus, rigorous performance verification is essential to ensure these barriers remain hydraulically effective after decades of operation.

The purpose of this study is to present a methodology for effectively diagnosing the condition of cut-off barriers in groundwater dams and to verify its validity through practical application cases. The subject of this study is a concrete vertical cut-off barrier with a crest width of less than 1 meter and a total length of approximately 800 meters. The location of the cut-off barrier was first identified using electrical resistivity survey, a non-destructive method; subsequently, borehole investigations and permeability tests were conducted to analyze the characteristics of the ground and the barrier materials. The groundwater flow characteristics were confirmed through groundwater flow direction and velocity measurements. Furthermore, the internal condition of the cut-off barrier and the surrounding ground were comprehensively evaluated by performing electrical resistivity tomography (ERT) using investigation boreholes installed upstream and downstream of the barrier.

Evaluation results revealed seawater intrusion at the barrier-natural ground boundary and localized deterioration, indicating a decline in seepage control. Such degradation threatens the storage efficiency and water supply reliability of the groundwater dam. Consequently, a systematic diagnosis framework is imperative for continuous operation. Beyond short-term remediation, it is essential to implement long-term repair and reinforcement strategies. This study provides the foundational data required to develop these management systems and serves as a model for maintaining other groundwater dams.

Acknowledgements This work was supported by the Management Technology for Groundwater Dams in Water Supply Vulnerable Areas Program of the Korea Environmental Industry & Technology Institute(KEITI), funded by the Ministry of Environment (MOE) (RS-2025-01842972).

ABSTRACT. Nutrient runoff derived from intensive agricultural practices in upland fields is transported to adjacent rivers and lakes during rainfall events, causing water quality degradation and eutrophication. Agricultural nutrient losses are influenced not only by nutrient input rates, topography, and management practices but also by rainfall characteristics such as precipitation and rainfall intensity. To reduce nutrient losses, various best management practices (BMPs), including no-tillage, cover crops, filter strips, and fertilizer management, have been widely adopted. However, as climate change alters rainfall patterns, the effectiveness of BMPs is expected to vary across farmlands with different hydrologic and management conditions. Despite this potential variability, quantitative assessments of climate change impacts on BMP effectiveness at the field scale remain limited. The objective of this study is to quantify changes in BMP effectiveness under climate change conditions and to evaluate variations in nutrient reduction among agricultural fields. We used the process-based Agricultural Policy/Environmental eXtender (APEX) model, which is well suited for simulating soil and water quality responses to agricultural activities at the field scale. The APEX model was constructed using monitoring data collected from a corn–autumn cabbage double-cropping field at the Kangwon National University field plot from April to October 2025. Observed surface runoff, total nitrogen (TN), and total phosphorus (TP) loads were used for model calibration. The calibrated APEX model was driven by historical climate observations and future climate scenarios based on Shared Socioeconomic Pathways (SSPs). Multiple BMP scenarios, including no-tillage, cover crops, filter strips, and fertilizer management, were implemented to simulate nutrient runoff under contrasting climate conditions. Model simulations were conducted to compare BMP effectiveness between historical and future climate scenarios and to assess changes in BMP effectiveness across different management practices. This study evaluates differences in BMP effectiveness under present and future climate conditions to identify potential shifts in nutrient reduction induced by climate change. The results provide insights into how BMP effectiveness may vary under changing climate conditions and support the development of adaptive BMP implementation strategies tailored to field-scale conditions. This study provides a scientific basis for evaluating BMP effectiveness under climate change and contributes to improving agricultural water quality under future climate conditions.

ABSTRACT. This study aimed to elucidate the impacts of meteorological variations and watershed environmental changes induced by climate change on the dynamics of the Metalimnetic Oxygen Minimum (MOM) layer in Lake Soyang, a large stratified reservoir in South Korea. To achieve this, the CE-QUAL-W2 model, a laterally averaged two-dimensional hydrodynamic and water quality model, was constructed. The model's high reproducibility was verified through a two-year simulation (2023-2024) based on observed data (Water Level R2=0.99, Water Temperature R2=0.97, Dissolved Oxygen R2=0.65). This study constructed multifaceted scenarios involving water temperature rise, wind speed variations, and fluctuations in inflow dissolved organic matter (DOC) loads to simulate potential extreme environmental conditions caused by climate change. The impact of each factor on the vertical structure of dissolved oxygen in the reservoir was quantitatively analyzed. Simulation results indicated that the MOM in Lake Soyang typically forms in mid-June and exhibits structural characteristics reaching maximum development between August and October at depths of 20-35 m. Furthermore, rising water temperatures and decreased wind speeds were found to enhance reservoir stratification stability by increasing the density difference with the hypolimnion and inhibiting physical mixing. In conclusion, future water quality management for Lake Soyang requires an integrated approach that combines dam operation strategies considering changes in physical stratification intensity due to climate variability with strategies to suppress biogeochemical oxygen consumption through the reduction of watershed DOC loads. These findings are expected to contribute to the establishment of dam and watershed management measures that account for the complex interactions of hydrodynamic and water quality factors in the era of the climate crisis.

ABSTRACT. Soyang Lake is one of the largest artificial lakes in Korea, located in the upstream of the North Han River system. Since a cyanobacterial bloom was first officially reported around Inje Bridge in 2023, blooms have occurred consecutively in 2024 and 2025. This study aimed to characterize phytoplankton community shifts associated with these blooms, identify dominant environmental factors, and examine the co-occurrence of the cyanotoxin microcystin-LR (MC-LR). Water samples were collected at five upstream sites—Livingston Bridge (LS), Naerincheon (NR), Salgumi Bridge (SG), Gunchuk Bridge (GC), and Golmal Bridge (GM)—once per month from June to October 2024 and from March to October 2025; during the cyanobacterial season (June–September), sampling was conducted at least twice per month. Phytoplankton were identified and enumerated by microscopy. To examine relationships between environmental factors and phytoplankton dynamics, water-quality variables (water temperature, pH, total nitrogen, total phosphorus, and chlorophyll-a) were measured, and daily meteorological data (air temperature and precipitation) from 2015 to 2025 were used. Community structure was evaluated using principal coordinates analysis (PCoA) and the Shannon and Gini–Simpson indices, and trophic status was assessed using the Korean Trophic State Index (TSIKO). MC-LR was examined in relation to periods and locations of cyanobacterial dominance. Across the study period, Cryptomonas (Cryptophyta) tended to dominate at all sites. At Gunchuk Bridge, however, the summer community shifted to cyanobacteria, with Dolichospermum and Microcystis jointly accounting for more than 50% of the assemblage. PCoA and diversity indices showed that the Livingston Bridge–Naerincheon–Salgumi Bridge reach maintained a relatively diverse assemblage in which diatoms, green algae, and cryptophytes co-occurred, whereas Golmal Bridge consistently exhibited low diversity and high dominance. At Gunchuk Bridge, bloom periods repeatedly coincided with near single-taxon dominance and a sharp decline in diversity. TSIKO indicated an overall mesotrophic condition on an annual-average basis, but Gunchuk Bridge temporarily increased to eutrophic in 2024 and hypertrophic in 2025, consistent with cyanobacterial mass occurrence. MC-LR was co-detected at times and locations with pronounced cyanobacterial dominance. Overall, phytoplankton community fluctuations differed among sites, and at Gunchuk Bridge, cyanobacterial dominance repeatedly occurred, accompanied by a decrease in species diversity, an increase in TSIKO, and the concurrent occurrence of MC-LR. These results show that the Gunchuk Bridge reach is a key vulnerable section where bloom occurrence and toxin risk are simultaneously concentrated, and indicate that upstream nutrient reduction and residence-condition management through hydrologic operation are required as priorities.

ABSTRACT. Riparian land purchase and the subsequent restoration of natural land cover are recognized as effective mechanisms for simultaneously mitigating non-point source pollution and organic carbon runoff. However, research systematically quantifying the correlation between land purchase and carbon (TOC/POC) export dynamics at the watershed scale in Korea remains limited. Therefore, this study applies the Soil and Water Assessment Tool-Carbon (SWAT-C) model to the upstream Heukcheon A sub-watershed, a tributary of the Namhan River within the Han River watershed. [EH2.1]The primary objective of this study is to quantify the reduction in organic carbon runoff resulting from cropland acquisition by establishing a watershed-scale analytical framework. To conduct this study, a database was constructed using spatial data (including DEM, land cover, and soil maps) and observed meteorological, hydrological, and water quality data. The reliability of the hydrological simulation was first ensured through the calibration and validation of streamflow using the SWAT model. Subsequently, the SWAT-C carbon module was integrated to optimize the simulation of carbon dynamics within the watershed by comparing and calibrating the model against observed TOC concentrations and loads. Furthermore, comparative simulations were performed between the baseline (pre-purchase) and scenarios converting riparian cropland into forest, grassland, and wetland. Key indicators, such as the annual average TOC reduction per unit area (ha), were calculated. By providing objective quantitative indicators of water quality improvement and carbon reduction through the validated SWAT-C framework, this study establishes scientific evidence to support land use strategies in drinking water management areas and the expansion of watershed-based carbon neutrality policies.

ABSTRACT. The Habitat Suitability Index (HSI) is a key metric for quantifying the ecological responses of fish species to environmental variables such as water depth, temperature, and velocity. Despite its importance, conventional HSI estimation often depends on manual curve selection and subjective parameterization, limiting reproducibility across basins and time periods. To overcome these challenges, we developed a Python- and Streamlit-based automated HSI estimation program capable of processing long-term, multi-basin datasets and producing standardized suitability curves. This study utilized long-term national fish monitoring data (2011–2024) collected by the National Institute of Environmental Research (NIER), where fish assemblage surveys are conducted twice annually at fixed national monitoring sites. The proposed program consists of three core modules: IFASG (Incremental Frequency Analysis of Species Group)-based distribution normalization that converts irregular species-count distributions into ecologically interpretable percentile-based HSI reference points; automatic fitting of seven nonlinear models (Weibull, Gamma, Logistic, Gompertz, Arctangent, Natural-growth, and Generalized Poisson); and quantitative model evaluation using R², Pearson correlation, and Spearman rank correlation, enabling objective selection of the optimal model for each species–habitat variable combination. The tool automatically ingests species-by-basin Excel datasets, aggregates long-term observation counts, constructs IFASG-based suitability profiles, and produces detailed visualizations, parameter tables, and suitability thresholds. Application to the Namhan and Bukhan River basins revealed clear inter-basin differences in depth-, temperature-, and velocity-dependent habitat responses. Given its ability to process long-term national monitoring data and standardize HSI derivation, the developed program has strong potential for supporting ecological flow assessments, habitat simulation (PHABSIM/IFIM), and basin-scale ecological evaluations. It also offers a reproducible and transparent framework for future fish-habitat decision-support systems.

ABSTRACT. Forest carbon accumulation reflects how vegetation growth translates into biomass under specific environmental and ecological conditions, yet its variation across species and site characteristics remains insufficiently quantified using observational data. In this study, forest carbon accumulation across South Korea is quantified by converting growing stock data from national forest test sites into biomass and carbon mass for coniferous, broadleaf, and mixed forests. The analysis focuses on how carbon accumulation varies across sites in relation to soil moisture, species composition, competition, soil properties, and land conditions. By directly comparing carbon estimates across different environmental and ecological settings, we examine how site conditions and forest structure influence biomass accumulation. This approach enables a data-driven assessment of how different species and environmental factors control the amount of carbon stored in forest biomass. The results provide a quantitative basis for understanding how forest growth is regulated by environmental constraints and offer insight into how carbon accumulation varies under different ecological conditions.

ABSTRACT. The algal ball Marimo (Aegagropila linnaei) is a unique spherical alga that can grow to a diameter exceeding 30 cm. Recognized as a special Japanese natural treasure in 1952, Marimo holds significant ecological importance. Lake Akan is designated as the only lake in the world where large Marimo form dense colonies. Despite various conservation efforts and scientific studies conducted to date, a critical gap remains: no indicator has yet been established to assess the health condition of Marimo, a tool essential for its effective protection and management. This study aims to develop an evaluation method for Marimo condition. To achieve this, we employed a novel dual-parameter approach. Firstly, we measured the amount of accumulated internal septic matter within the Marimo. Secondly, we conducted an underwater drop experiment to examine the organism's elastic response, enabling the calculation of its elastic modulus. The results showed a clear trend: the weight of the septic matter increased as the diameter of the Marimo increased. Simultaneously, the elastic modulus also rose with increasing diameter, indicating that larger amounts of septic matter generally corresponded to higher elastic modulus values. Crucially, however, when comparing Marimo of the same size, those with less septic matter exhibited a higher elastic modulus. This suggests that Marimo with lower septic matter content may be in a healthier condition. Furthermore, samples were collected from two distinct areas of Lake Akan: Churui Bay and Kinetampe Bay. Marimo from Churui Bay contained more septic matter, indicating a poorer condition. MRI imaging was used to provide insight into the internal structure. This confirmed the poor condition of the Churui Bay samples, which displayed internal cracks. 2-3 cm masses of organic material, previously interpreted as signs of decay in other studies, were also observed, suggesting an additional cause of deterioration. The correlation established between the quantity of internal septic matter and the elastic modulus provides a foundational step towards an objective and quantitative indicator for assessing Marimo health, which can significantly enhance conservation and management efforts.

ABSTRACT. The phenomenon of total nitrogen exceeding the standard in the water body of the Yellow River Basin had not been effectively solved yet，From 2024 to 2025, the average total nitrogen (TN) concentration in water bodies of large-scale reservoirs along the main stream of the Yellow River exceeded 2 mg/L, which posed risks of environmental pollution and adverse impacts on human health. Sediment was an important medium for total nitrogen land-water migration. The loess area in the Yellow River Basin had a large distribution area, poor erosion resistance, and high total nitrogen flux into the river. Therefore, taking the typical small watershed in the shallow mountainous area of the Loess Plateau as the test area, the total nitrogen interception efficiency of polyploid Arundo donax was evaluated by the method of runoff plot test under natural rainfall in the field, and the value of its ecological products was evaluated by the ecological value method. The results showed that the total nitrogen interception efficiency of fertilized grass in the first year of planting was 89.41 % and 29.70 % higher than that of native high-coverage herbaceous plants and arbor-herbaceous plants, respectively. The annual ecological product value per hectare was 141.3 million yuan, of which the supply value of agricultural products was 79875 yuan, the soil conservation value was 1618 yuan, the carbon sequestration value was 453 yuan, the oxygen release value was 1972 yuan, the water conservation value was 21400 yuan, and the employment service value was 36,000 yuan. In conclusion, polyploid Arundo donax exhibited high efficiency in pollutant interception and economic conversion value, and thus could be recommended as a priority species for soil and water loss control and riparian buffer zone construction in Yellow River Basin.

ABSTRACT. Laundry activities in residential and commercial sectors are a major source of urban water consumption and wastewater generation, with the rinsing stage in particular involving substantial water use and pollutant discharge. This study aims to quantitatively analyze how a reduction in rinsing frequency enabled by the adoption of surfactant-free detergents can be scaled up to city-level reductions in water demand, energy consumption for water and wastewater treatment, and associated carbon emissions. The study area is Seoul, South Korea, encompassing all households and commercial facilities served by the municipal water supply and sewerage systems, as well as water and wastewater treatment plants. The analysis follows a structured framework consisting of scenario development, estimation of water savings, calculation of energy savings and carbon emission reductions, and integrated economic and environmental assessment. The baseline scenario assumes an average of two rinsing cycles per laundry event with zero adoption of eco-friendly detergents, while alternative scenarios consider a reduced rinsing frequency (from two to one) combined with detergent adoption rates ranging from 30% to 100%. The results quantitatively demonstrate that a reduction in rinsing frequency as a daily behavioral change can be translated into measurable reductions in urban water use, as well as energy consumption and carbon emissions associated with water and wastewater treatment. These findings indicate that the adoption of eco-friendly consumer products can contribute to improving the efficiency of the urban water–energy–carbon nexus, and provide a quantitative basis for supporting related policy development and sustainable urban management strategies.

ABSTRACT. The necessity for a stable and efficient operation of wastewater treatment plants is increasing, as urbanisation and population growth lead to an increase in both influent flow rate and water quality variability. However, the existing method of manual process control based on the controller’s experience may trigger energy overuse and fluctuations in effluent quality. To address these issues, the study developed a cyber-physical system (CPS) integrating artificial intelligence for the MLE process and verified its performance through application to a 100m³/d MLE Pilot Plant. The wastewater treatment process with CPS applied can provide optimal operating conditions by collecting real-time process data and autonomously adjusting aeration and return flow rates. The results indicated that approximately a 10% reduction in air flow rate has brought an improvement in energy efficiency while maintaining stable effluent quality. Furthermore, by applying CPS to practical wastewater treatment plants, accumulating long-term operational data, and advancing control algorithms based on the accumulated data, this study is expected to provide foundational evidence for the smart transformation of wastewater treatment plants and sustainable water resource management.

ABSTRACT. At river confluences, tributaries carry different temperatures, dissolved ions, and flow energies, creating complex mixing patterns that evolve rapidly in space and time. Traditional source tracing techniques, such as isotope and discrete sampling, provide reliable chemical identification; however, high costs, low temporal resolution, and limited spatial coverage constrain their application. They are also difficult to utilize for real-time monitoring after rainfall or reservoir releases, or during rapid hydrological changes. This study introduces a new framework that adapts a widely used method in oceanographic analysis. In oceanography, the T-S diagram (water temperature-salinity) technique can be used to track the origin of water bodies and analyze the mixing process by using the correlation between water temperature and salinity. Most rivers in Korea are characterized by freshwater environments. The temperature–salinity diagram for freshwater systems can be modified by replacing salinity with electrical conductivity, a variable strongly linked to the concentration of dissolved ions. The resulting Temperature–Conductivity (T–C) diagram enables the visualization of water mass origins, mixing trajectories, and density-controlled interactions without the need for intensive laboratory analysis. We applied this approach to the confluence integrating high-resolution boat-based measurements with hourly data from the national monitoring network. Based on integrated data, we compared vertical profile data from field measurements and T-C diagrams. Further, we presented spatial interpolation contour data based on the measured data, demonstrating that the practical application of T-C diagrams is feasible. Results show that the T-C framework successfully distinguished non-mixed and partially mixed samples and was also able to distinguish fully equilibrated flow regions, including seasonal and rainfall-induced transitions. It appears that the t-c diagram can also be used to partially evaluate the flow of diatoms or green algae.

This work was supported by the Ministry of Climate, Energy, Environment through the project 'Research and Development on Technology for Securing Water Resources Stability in Response to Future Changes (grant number RS-2024-00332114)'.

ABSTRACT. Urban flooding has become increasingly frequent and severe due to climate change and rapid urbanization, highlighting the limitations of conventional river-centered flood control measures. In this context, detention basins connected to river systems have been emphasized as an effective distributed strategy for urban flood mitigation. However, the hydraulic performance of detention basins, particularly side-weir overflow behavior under extreme flow conditions, remains insufficiently verified through experimental evidence. The objective of this study is to experimentally evaluate the flood mitigation effects of a detention basin side-weir under urban flood conditions. Pilot-scale physical experiments were conducted at the Andong River Experiment Center, where a detention basin system including a side-weir was constructed to reproduce realistic river–detention interactions. Stepwise gate opening scenarios were applied to simulate increasing main-channel discharge, enabling controlled reproduction of detention inflow, overflow initiation, and storage operation processes. Hydraulic responses were comprehensively monitored using multiple instruments, including acoustic Doppler current profilers, radar and pressure-based water level gauges, electromagnetic surface velocity meters, and unmanned aerial vehicle imagery. The integration of in-situ measurements and image-based velocity estimation allowed detailed analysis of spatial and temporal hydraulic behavior around the side-weir and detention basin. Experimental results indicate that detention basin operation effectively attenuates water level rise in the main channel during increasing discharge conditions. In particular, flow diversion through the side-weir significantly contributes to peak water level reduction, demonstrating the role of side-weir overflow as a key mechanism for urban flood mitigation. The study also establishes an experimental database of side-weir hydraulic characteristics derived from multi-sensor observations. The findings provide practical insights for the design and operation of detention basins linked to urban rivers and support the development of resilient flood mitigation strategies under increasing climate-induced flood risks.

This work was supported by Korea Environment Industry & Technology Institute(KEITI) through Climate Resilient R&D Project for Water-Related Disaster Management, funded by Korea Ministry of Climate, Energy and Environment(RS-2023-00218973)

ABSTRACT. This study proposes an impact-based quantitative assessment framework to objectively identify manholes that exert significant influence on underground flooding within urban drainage systems, with a focus on the area surrounding Seoul National University of Education Station in Seoul, South Korea. An integrated modeling framework coupling the Storm Water Management Model with a two-dimensional surface inundation model was developed to simulate multiple rainfall scenarios with return periods ranging from 30 to 500 years and durations between 60 and 180 minutes. Key evaluation indicators included the spatial distance between manholes and subway entrances, the timing of backflow occurrence within the sewer network, and cumulative manhole overflow volumes. The results indicate that specific manholes (8201, 2112, and 4100) consistently exhibited high impact scores under all simulated conditions, playing a critical role in the propagation of underground flooding. The proposed methodology provides a quantitative basis to supplement conventional experience-based management practices and is expected to support more rational decision-making for underground flood early warning systems and priority manhole management.

Funding This work was supported by the Technology Innovation Program (RS202400398858, Development of AI-based urban flood damage risk prediction and evaluation technology for practical use) funded By the Ministry of the Interior and Safety(MOIS, Korea), and by National Disaster Management Research Institute (NDMI), Ministry of the Interior and Safety, Republic of Korea (Project No. RS-2025-02218717).

ABSTRACT. Satellite-based precipitation products provide an opportunity to complement ground observations and enhance the spatial characterisation of rainfall fields. This study presents a methodological framework for integrating satellite-derived precipitation data with ground-based observations in an instrumented region of Sicily (Italy) to assess how satellite data integration affects the estimation of parameters used to derive Intensity–Duration–Frequency (IDF) curves. The methodology is applied to the Sicily region (Italy), covering approximately 25,800 km². Satellite rainfall data were obtained from the Global Precipitation Measurement (GPM) Multi-Satellite Precipitation Analysis product, which provides precipitation estimates with a spatial resolution of 0.1° × 0.1° and a temporal resolution of 30 minutes. These data were integrated with ground observations from the SIAS (Servizio Informativo Agrometeorologico Siciliano) monitoring network, comprising 108 rain gauge stations that provide rainfall measurements at a 10-minute temporal resolution. Rainfall datasets were first subjected to quality control and preprocessing procedures. Missing observations in the rain gauge series were reconstructed using a stochastic rainfall generator based on the Bartlett–Lewis model. Subsequently, rainfall data were aggregated to a 30-minute time step in order to ensure consistency with the satellite observations and allow a direct comparison between the two datasets. This step enabled the spatial calibration of satellite rainfall estimates using rain gauge observations in order to correct systematic biases and improve the representation of spatial rainfall variability. Following this step, the temporal calibration of the satellite precipitation data was carried out through a downscaling procedure based on high-resolution (10-minute) rain gauge measurements, ensuring consistency with the observed rainfall dynamics. After the calibration phase, IDF curves were derived for each monitoring station and the parameters of the intensity–duration relationship were estimated. Spatial interpolation of the parameters aand nwas then performed using kriging to generate regional maps describing their spatial variability. To assess the added value of satellite data, two modelling configurations were considered: a reference model based exclusively on ground observations and an integrated model combining rain gauge and satellite precipitation datasets. The spatial patterns of the IDF parameters obtained from the two approaches were compared to evaluate the impact of satellite data integration on the representation of extremes. The inclusion of satellite information improves the spatial continuity of precipitation fields and provides a more robust characterisation of rainfall extremes. The proposed framework demonstrates the potential of integrating satellite and ground-based observations to enhance the reliability of IDF relationships, thereby supporting improved hydrological assessments and more robust hydraulic design.

ABSTRACT. Urban flood risk caused by pluvial rainfall has intensified under the combined influence of climate change and rapid urbanization, exposing fundamental limitations in conventional flood management strategies. Most existing approaches rely on fixed measures such as infiltration facilities or underground detention systems. While these systems have played an important role, their effectiveness is increasingly constrained by limited land availability, difficulties in large-scale expansion, and reduced adaptability to highly localized and extreme rainfall patterns.

This study introduces spatial detention technology as an alternative approach that broadens the role of urban surface space in flood management. Rather than treating detention as a permanent structure, spatial detention views certain urban spaces as resources that can temporarily function as storage areas during flood events. By coordinating movable flood barriers with surface topography, drainage networks, and hydrological conditions, spaces such as roads, plazas, and underutilized open areas can be selectively utilized to store excess runoff when needed. This allows flood mitigation functions to be distributed across the city while maintaining normal urban use under non-flood conditions.

The paper reviews recent research and technical developments related to spatial detention technology, focusing on how the concept has been structured and applied in practice. Attention is given to the engineering definition of spatial detention and its physical limits, the identification of suitable urban spaces based on hydraulic and spatial characteristics, and the development of operational scenarios that link movable structures with surface runoff behavior and drainage performance. Through this discussion, the study highlights a shift away from static, design-based flood control toward a more flexible, scenario-driven response framework.

Future work will examine application cases and pilot-scale demonstrations to quantitatively assess the potential of spatial detention in reducing peak runoff, delaying flow concentration, and redistributing inundation risk at the urban scale. These investigations are also expected to clarify the practical strengths of the approach, including operational flexibility, phased implementation, and compatibility with existing urban infrastructure. Overall, the study positions spatial detention technology as a realistic and scalable option for urban flood management under increasing uncertainty, and as a step toward a more adaptive, space-oriented flood control paradigm.

ABSTRACT. This study presents a practical evaluation framework that integrates a Flood Defense Index (FDI) with the four resilience components (4R): Robustness, Rapidity, Redundancy, and Resourcefulness. The FDI is quantified at the conduit scale as the ratio of simulated flow depth to pipe diameter (d/D), providing a high-resolution spatial and temporal assessment of hydraulic loading and system performance.The framework was implemented for the Gunja and Banpocheon watersheds in Seoul using EPA SWMM 5.2. To ensure a robust stress test, Huff third-quartile design rainfall was applied across various return periods and durations. We compared three infrastructure configurations: (i) Sewer Pipes only (SP), (ii) SP with Detention Basins (SP–DB), and (iii) SP–DB with Pumping Stations (SP–DB–PS). Results demonstrate that the SP–DB–PS scenario delivers the most stable flood-defense performance. While the SP-only scenario suffered from sharp FDI declines under high-intensity events, the addition of detention storage effectively dampened peak responses, and pumping stations further enhanced the system’s ability to maintain functional capacity. Validation using historical observed rainfall events confirmed consistent scenario rankings and 4R tendencies, proving the framework’s reliability. The proposed FDI–4R approach provides a quantifiable and intuitive metrics-based tool for urban planners to optimize drainage infrastructure and enhance climate change adaptation strategies.

ABSTRACT. Thornthwaite and Mather daily water balance model, as defined by the WRP12 guideline in Malaysia, is widely used for applications such as water availability assessment for ungauged or data-scarce catchments. The model is simple and involves minimal input parameters, which makes it suitable for preliminary analysis. An R script for the calculation of this water balance model was developed. The efficiency of the R script in replicating the workflow for Thornthwaite and Mather daily water balance model and its ability to generate the results significantly faster than a spreadsheet proves that the script-based approach provides a streamlined alternative for such tasks, besides ensuring reproducibility and transparency. The water balance model is applied for the assessment of surface water resources in Peninsular Malaysia. There is potential for extending the use of the model to include the assessment of other water balance components such as groundwater recharge.

ABSTRACT. Accurate runoff estimation is fundamental to the design of hydraulic structures and the mitigation of inland flooding in urban areas. This issue is particularly critical in small-scale watersheds, where the Rational Method is commonly applied, making the accuracy of rainfall intensity—a key determinant of runoff magnitude—particularly important. However, most existing intensity–duration–frequency (IDF) curves have been developed based on rainfall data with temporal resolutions of 10 minutes or longer. This limitation has led to persistent accuracy issues, particularly when the time of concentration is less than 10 minutes, because such events are typically constrained by the minimum 10-minute duration. Although previous studies relied on empirical conversion models or statistical estimation due to the limited availability of high-quality 1-minute rainfall data, recent advancements in high-resolution rainfall observation technologies now enable the direct use of long-term, fine-scale time-series datasets. This study develops short-duration IDF curves using 1-minute rainfall data from 22 Automatic Weather Stations (AWS) in Seoul, each providing more than 25 years of continuous observations. Annual Maximum Series (AMS) were constructed for multiple rainfall durations, followed by frequency analysis. A sixth-order log-polynomial equation was employed to enhance the goodness of fit of the resulting IDF curves. The results indicate a significant reduction in the distortion of rainfall intensity estimates typically observed at short durations (10–60 minutes), leading to improved accuracy in the estimation of design rainfall intensity and depth. Furthermore, the high spatial density of the AWS network—more than twice that of the Automated Surface Observing System (ASOS)—allows for a more detailed representation of rainfall spatial variability. These findings provide a robust technical basis for establishing urban disaster prevention design standards that adequately reflect sub-10-minute rainfall characteristics.

ABSTRACT. Due to climate change, the frequency of localized heavy rainfall and extreme precipitation has increased, leading to a growing risk of urban flooding in areas with high impervious surface ratios. As a result, urban areas with dense transportation, industrial and residential functions experience compounded economic losses and casualties during flood events. To reduce these risks, short-term flood prediction technologies that can reflect real-time rainfall characteristics are required. This study proposes a short-term urban flood prediction approach that couples a U-Net based deep learning rainfall prediction model(Recursive RainNet) with an inundation prediction model(CRU-Net). Recursive RainNet was trained using S-band weather radar gridded rainfall data to predict rainfall distribution with a lead time of up to one hour. CRU-Net was trained using physics-based flood scenarios incorporating topographic and urban runoff characteristics, and utilized predicted rainfall as input to estimate inundation depth at 10-minute intervals for up to 60 minutes. The results indicated that Recursive RainNet effectively captured the spatiotemporal variability of rainfall and demonstrated stable short-term rainfall prediction performance. In addition, the inundation predictions generated by CRU-Net spatially matched observed flood maps, successfully reproducing actual flood-prone areas. This study evaluated the feasibility of real-time urban flood prediction by applying coupled deep learning-based rainfall and inundation models to a real urban case. The proposed approach is significant in that it enables proactive prediction of flood occurrence even under highly variable localized heavy rainfall conditions, demonstrating its potential for use in early warning and decision support system for urban flood management.

Acknowledgements

The research for this paper was carried out under the KICT Research Program (Project no. 20250284–001, Development of Digital Urban Flood Control Technology for the Realization of Flood Safety City) funded by the Ministry of Science and ICT.

ABSTRACT. Urban flooding causes serious economic damage in cities with high population and asset densities. While physics-based models have traditionally been used to predict flood susceptibility by simulating hydrological processes using hydraulic equations, machine learning (ML) models have recently gained attention for their ability to efficiently analyze high-dimensional data. However, urban infrastructure, which plays a critical role in urban flooding, has often been overlooked in flood-susceptibility mapping. Therefore, this study investigates how urban infrastructure variables, including road density, sewer system density, and urban detention water system density, influence the predictive performance of ML-based urban flood susceptibility models in Seoul, South Korea. Two input datasets were constructed by including or excluding urban infrastructure variables, while retaining the same topographic and pedological factors. Flood inventory data derived from historical flood inundation maps (2010–2024, excluding 2015 and 2021) were used to label flooded and unflooded areas. To address class imbalance, 1,000 flooded and 1,000 unflooded points were randomly sampled, yielding 2,000 samples. The sampling was repeated five times to enhance robustness. Each dataset was divided into training (70%) and validation (30%) subsets. Random forest (RF), extreme gradient boosting (XGB), multi-layer perceptron (MLP), long short-term memory (LSTM), and a transformer-based model (TabPFN) were applied. Hyperparameters for RF, XGB, MLP, and LSTM were optimized using Bayesian optimization, whereas TabPFN was applied without additional tuning. Model performance was evaluated using confusion-matrix-based metrics, including accuracy and ROC–AUC. SHAP analysis was employed to identify key contributing variables. Additionally, the mapping results were divided into 5 grades (very-low~very-high), and the very-high grades were compared spatially with flood inventory data, using recall and f1-score with urban variables. The results showed that ML models with urban infrastructure achieved ~0.03 higher accuracy and ROC-AUC than those without urban infrastructure. Among ML models, TabPFN achieved the highest performance, with an average accuracy of 0.81 and an ROC–AUC of 0.90. Road density was the most influential variable, followed by elevation. Additionally, urban infrastructure improved the spatial detectability of urban flood susceptibility in the mapping results and agreement with historical flood records. These findings suggest the importance of urban infrastructure and the effectiveness of a transformer-based model for urban flood susceptibility, supporting ML-based flood management and urban flood prevention strategies.

ABSTRACT. Major metropolitan areas increasingly adopt deep underground stormwater tunnels as a key measure for urban flood mitigation. Previous studies on individual vertical shafts have primarily focused on geyser phenomena and their association with water-hammer effects and reverse-flow behavior. Under initial inflow conditions with low downstream water levels, such drainage systems operate in a free-surface flow regime. As extreme rainfall events raise downstream water levels, the flow gradually transitions to a pressurized state, during which air–water two-phase flow develops and air pockets become entrapped within the tunnel. This transition from free-surface to pressurized flow induces reverse flow. When the resulting backward surge reaches the downstream outlet, a rapid pressure rise caused by increased water levels and sudden flow deceleration generates a strong reverse surge. The entrainment of air within this reverse flow produces significant pressure fluctuations throughout the tunnel, potentially compromising structural safety and operational stability. In this study, pressure characteristics at the bottom junction of a vertical inlet shaft in an underground drainage tunnel are experimentally investigated, with particular attention to energy dissipation associated with flow behavior and the installation of a water cushion. Laboratory-scale experiments are conducted using a physical flume model consisting of an inflow vertical shaft, a drainage tunnel, ventilation shafts, an outflow shaft, a surge tank, and a collection tank. The experimental system has a total length of approximately 15.3 m and a maximum height of about 5 m, representing a scaled configuration of a typical deep underground drainage tunnel. Three ventilation shafts are installed along the drainage tunnel, and their locations can be adjusted according to the experimental configuration. Backflow conditions within the tunnel are controlled by regulating the valves of the surge tank and the collection tank, allowing systematic observation of pressure response and flow behavior under varying hydraulic conditions.

ABSTRACT. Urban pluvial flooding driven by short-duration intense rainfall has become more frequent in recent years. To reduce flood risk, deep underground drainage tunnels with large storage capacity have been widely adopted in urban areas. Nevertheless, overflow events are still observed even under similar rainfall conditions, suggesting that factors beyond design rainfall and peak inflow may influence system performance.

This study explores the role of initial storage conditions in overflow occurrence within a deep urban drainage tunnel system. A probabilistic model is applied to examine the relationship between initial storage level, inflow characteristics, and overflow occurrence across multiple vertical shafts. Model performance is evaluated using standard probabilistic metrics, and differences in overflow behavior among shafts are examined with respect to overflow duration.

The results indicate that initial storage conditions have a noticeable influence on overflow occurrence, while structural differences among vertical shafts contribute to variations in overflow persistence. These findings highlight the importance of considering pre-event system conditions when interpreting overflow behavior in large underground drainage systems.

This work was supported by Korea Environment Industry & Technology Institute(KEITI) through Climate Resilient R&D Project for Water-Related Disaster Management, funded by Korea Ministry of Climate, Energy and Environment(RS-2023-00218973)

ABSTRACT. Urban flooding has intensified in densely populated cities due to climate change, increased impervious surfaces, and limitations of urban drainage systems. Accurate and interpretable flood susceptibility assessment is therefore essential for effective urban flood management. This study develops an explainable artificial intelligence (XAI)-based machine learning framework for high-resolution urban flood susceptibility prediction.

Four machine learning models—Artificial Neural Network (ANN), Support Vector Machine (SVM), Random Forest (RF), and Extreme Gradient Boosting (XGBoost)—were applied to Seoul, South Korea. A spatial database was constructed at a 5 m grid resolution using geographic information system (GIS) techniques, incorporating eight flood-influencing factors: elevation, slope, aspect, topographic wetness index, impervious surface ratio, distance to rivers, distance to drainage conduits, and distance to manholes. Historical flood inventory data from 2010 to 2023 were converted into point-based datasets, and 1,000 flood and non-flood locations were selected to construct balanced training and testing datasets.

Model performance was evaluated using the area under the receiver operating characteristic curve (AUC) and the Kappa coefficient. XGBoost showed the highest predictive accuracy, while Random Forest demonstrated greater robustness and spatial consistency for flood susceptibility mapping. Flood susceptibility maps were produced and classified into five levels using the Jenks natural breaks method, revealing clear spatial variability across urban districts.

To improve model interpretability, SHapley Additive exPlanations (SHAP) were applied to quantify the contribution of individual factors. A Borda count aggregation of SHAP results identified impervious surface ratio as the most influential factor, whereas slope had the least impact. The proposed XAI-based framework provides a transparent and reliable tool for urban flood susceptibility assessment and supports evidence-based non-structural flood mitigation and policy decision-making.

This work was supported by Korea Environment Industry & Technology Institute (KEITI) through R&D Program for Innovative Flood Protection Technologies against Climate Crisis Project, funded by Korea Ministry of Environment (MOE) (2480000599)

ABSTRACT. Malaysia's rapid urbanization is projected to see over 80% of the population living in urban areas by 2040. This, along with climate extremes, has led to more frequent and severe urban flooding, causing significant losses in life, livelihoods, and property. This situation calls for urgent action to understand flood risk, enhance disaster governance, and strengthen community resilience, enabling urban populations to better prepare for future flood events. This study aims to provide new insights into developing an empirically grounded, policy-relevant framework to enhance community resilience in flood-prone urban environments. A survey was conducted among individuals directly or indirectly affected by the flood in Greater Kuala Lumpur, a highly urbanized region. The survey was analyzed using SEM AMOS, resulting in a higher order structure comprising nine constructs synthesized into three adaptive capacity domains: governance, responsive learning and community support. All three second-order constructs demonstrated robust measurement properties and strong support for the overall adaptive capacity construct. To contextualize these results and address institutional and behavioral dimensions, semi-structured interviews were conducted with key agencies and non-governmental organizations involved in urban flood management. The qualitative data were analyzed using NVivo, adopting an inductive thematic analysis approach to identify emergent themes related to adaptive capacity domains that explain how institutional coordination, knowledge dissemination and social relationships influence adaptive capacity in the urban flood context. Triangulation of these quantitative and qualitative results enabled the integration and corroboration of the findings, culminating in a final adaptive capacity framework that captures the interdependent contributions of key domains in shaping urban flood adaptive capacity. The study enhances urban flood resilience research by proposing an adaptive capacity framework that merges quantitative modelling with qualitative stakeholder insights. Key findings indicate that effective community responses to floods require aligned governance, ongoing capacity building, and strong social connections. This framework provides valuable guidance for disaster risk reduction and supports inclusive flood preparedness strategies in rapidly urbanizing areas, in line with the Sendai Framework, the IPCC AR6 Synthesis Report, and the 13th Malaysia Plan 2026-2030.

ABSTRACT. Rapid and largely unplanned urban fringe expansion in the Upper Selbe River area of Ulaanbaatar has increased exposure to flood hazards while weakening the ecological and hydraulic functions of the river corridor. In such peri-urban areas, absence of land-use zoning and regulation, encroachment into flood-prone zones, absence of drainage system, and the limited integration of water considerations into urban planning have intensified flood risk and reduced urban resilience. This study aims to assess flood risk in the unplanned urban fringe area of the Upper Selbe River and to identify the requirements for applying water sensitive urban design (WSUD) principles under the local environmental, hydrological, and urban development context of Ulaanbaatar, Mongolia. The research integrated hydrological and hydraulic modeling for flood assessment with spatial analysis of land use change, climate change impact, existing urban settlement patterns, drainage conditions, and river–floodplain interactions. The results showed that flood hazard in the Upper Selbe River fringe area is strongly influenced by titled settlement expansion into flood-prone zones, insufficient drainage system, reduced infiltration, and disruption of natural floodplain functions. Flood-prone zones were identified based on inundation extent and the exposure of vulnerable land uses and settlements. The findings indicate that effective flood risk reduction in the study area requires the application of WSUD measures adapted to local conditions, including retentions, infiltration enhancement, green corridor and wetland preservation, flood-compatible land use control, and river-sensitive spatial planning. The study concludes that integrating flood risk assessment with water sensitive urban design provides a practical basis for improving urban resilience, strengthening planning and regulatory frameworks, and advancing context-specific approaches to sustainable suburban development in cold-climate and data-constrained cities such as Ulaanbaatar.

ABSTRACT. Deep stormwater drainage tunnels have become a critical component of urban flood-mitigation systems in densely developed cities, where surface storage and conventional drainage capacity are limited. Under climate change, the increasing frequency and intensity of short-duration extreme rainfall events subject these deep conveyance systems to rapid inflow surges, generating strongly transient hydraulic conditions. Due to their substantial depth, confined geometry, and long hydraulic residence times, deep tunnels are particularly vulnerable to pressure transients driven by complex air–water interactions. Rapid inflow through dropshafts and junctions entrains large volumes of air into the tunnel system. The subsequent compression, migration, and release of entrapped air can induce severe transient phenomena, including pressure spikes, intermittent slug flow, and propagating pressurization waves. Although previous studies have advanced the understanding of air–water interactions in deep drainage tunnels, key uncertainties remain regarding the role of ventilation-shaft design in pressure stabilization under transient conditions. Existing research has largely relied on numerical simulations or small-scale laboratory experiments, which often struggle to reproduce large-scale air entrainment, compressive air-pocket dynamics, and pressure-wave propagation characteristic of full-scale tunnels. In particular, the combined effects of ventilation-shaft location, height, and number on transient pressure behavior have not been systematically validated through large-scale physical experiments. The present study conducts a controlled experimental investigation using a 15.3-m large-scale physical model designed to replicate realistic air–water interactions in deep stormwater drainage tunnels. The primary objective is to quantify how ventilation-shaft layout—specifically shaft location, height, and number—influences pressure stabilization during rapid inflow transients. By systematically varying these design parameters under identical hydraulic conditions, the study evaluates their individual and combined effects on air-pocket dynamics, slug-flow development, pressure oscillation characteristics, and overall ventilation performance. The findings provide experimentally validated insights to support safer tunnel design and improved operational reliability of deep drainage systems under extreme rainfall events. This work is financially supported by Korea Ministry of Climate, Energy, Environment(MCEE) as 「Climate Resilient R&D Project for Water-Related Disaster Management)(RS-2024-00397821)」This work is financially supported by Korea Ministry of Climate, Energy, Environment(MCEE) as 「Climate Resilient R&D Project for Water-Related Disaster Management)(RS-2024-00397821)」

ABSTRACT. Side weirs function as essential infrastructure for urban flood mitigation and flow regulation within open channel systems. Notwithstanding their importance, established empirical equations frequently yield inconsistent results across diverse hydraulic regimes. To address these limitations, this research proposes a refined empirical formula derived from high fidelity experimental datasets. By implementing a load cell based gravimetric system, we attained superior precision in discharge measurements compared to conventional approaches. The study investigates the interdependence between the discharge coefficient and governing variables, specifically the Froude number and geometric ratios (crest height and weir length). Our analysis demonstrates that models relying exclusively on the Froude number tend to underestimate discharge, particularly as weir length increases. By integrating both kinematic and structural parameters through regression analysis, the developed formula significantly enhances predictive robustness with a coefficient of determination of 0.97. These findings provide a more reliable framework for hydraulic design and analysis in complex flow environments.

ABSTRACT. Urban underground infrastructures are highly vulnerable to extreme flooding events, which are increasingly exacerbated by climate change and pose significant risks to densely populated areas. Motivated by the observation that existing research on flooded underground spaces is limited by underestimating the effects of human motion on evacuation safety, coarse flood simulations, fragmented risk assessments, and insufficient resilience management. To address these gaps, this study develops an agent-based dynamic vulnerability model for pedestrians in floodwaters, incorporating group interactions, formation patterns, and hydrodynamic forces, and supported by a full-scale instrumented physical model to capture interactive and dynamic evacuation behaviors. Analysis of spatial and temporal dynamics of pedestrian movement reveals significant variations in stability: walking against the flow increases instability and overall vulnerability, whereas moving with the flow reduces hydrodynamic forces, though this effect diminishes with increasing water depth. In parallel, a multi-dimensional hydrodynamic coupled model for urban water systems is established to enable high-precision simulations of urban runoff and drainage system flow interaction processes, accurately capturing the three-dimensional spatiotemporal evolution of flood intrusion into underground spaces. Building on these models, a comprehensive risk assessment framework is constructed by integrating multi-level evaluation metrics with scenario-based simulations, enabling effective quantification of risk propagation pathways within complex urban water systems. Results indicate that evacuation strategies in underground spaces should account for group-specific differences and the influence of the built environment, identifying critical evacuation times, implementing differentiated strategies, and managing congestion points to ensure efficient and safe passenger evacuation. This framework has been successfully applied in China-UK collaborative studies on urban flood prevention under extreme rainfall events. Overall, the framework provides practical guidance for designing flood-resilient underground spaces and delivers robust scientific and technical support for managing flood risks and enhancing urban resilience in representative urban underground infrastructures.

ABSTRACT. Small urban watersheds exhibit rapidly evolving hydrological responses over short time scales, in which water-level fluctuations are governed not only by total rainfall amount but also by the spatial distribution and intensity of rainfall relative to catchment boundaries. Even minor variations in rainfall can significantly affect the timing and magnitude of water-level rise and flood peaks. These dynamics highlight the need for short-term water-level forecasting to capture rapid responses. However, existing studies focus on single-station forecasting approaches that primarily learn localized temporal dependencies, limiting their ability to represent event-scale temporal context and inter-station connectivity in urban watersheds. This study proposes a hybrid Transformer–Bidirectional Long Short-Term Memory Model (Transformer-BiLSTM) for multi-station short-term water-level forecasting using minute-resolution data. The framework employs a Transformer encoder with learnable positional embeddings to capture long-range temporal dependencies across flood events, while a BiLSTM branch enhances learning of localized water-level variations and temporal continuity. By fusing global temporal representations with local sequential features, the model achieves balanced learning of hydrological dynamics across stations. Flood events over 11 years in the Kanda Basin, Tokyo, are used to develop and evaluate a model that learns from the preceding 60 minutes of rainfall data from 23 gauges and water-level data from 23 stations to forecast the upcoming 60-minute horizon at 23 water-level stations distributed across the basin. Three multi-station learning strategies: direct, station-specific, and group-based learning, and two forecasting approaches: single-stage sequence-to-sequence and two-stage recursive forecasting, are evaluated to verify the influence of spatiotemporal structuring on forecasting accuracy. Results demonstrate that the proposed models consistently outperform the baseline BiLSTM model, achieving median NSE above 0.9 at short lead times (≤ 30 minutes) and around 0.7 at longer lead times (40–60 minutes), with RMSE below 0.2 m. Downstream stations are generally predicted more accurately due to cumulative flow effects, whereas upstream stations exhibit greater uncertainty. At longer lead times (40–60 minutes), group learning with two-stage forecasting improves performance in the Zenpukuji watershed, while direct learning is more effective in the Kanda watershed, reflecting differences in spatial coherence and heterogeneity. These findings demonstrate that robust multi-station forecasting in small urban watersheds requires a balanced representation of temporal dependency, spatial hydrological connectivity, and rainfall characteristics, and that optimal model configurations depend on watershed structure and forecasting horizon.

ABSTRACT. Real-time urban flood forecasting requires fast and repeatable simulations that can be synchronized with evolving hydrology and meteorology observations. However, geometric reflection of building shapes within a computing mesh leads to significant increases in preprocessing and computational cost, which limits application feasibility. This study evaluates an urban flooding modeling approach that treats building effects as a grid-scale roughness upscaling problem by integrating hydraulic resistances through the roughness coefficient of composite manning rather than a geometric obstacle. Oncheon drainage basin, Pusan, Korea which was severely damaged by extreme rainfall events in 2024, was selected as a TestBed. Through manhole-based flux exchange, we applied an integrated double drainage framework (KICT-UF) that combines a 1-D sewer network solver and a 2-D surface flow solver. Surface routing was performed using a diffused wave format. Building induction resistance was implemented through a parameterized composite roughness formula, and the effective manning coefficient increases with the building coverage ratio, a local hydraulic head, and a building width promotion within each cell. Cells with building coverage exceeding 99% were treated as hydraulically blocked to ensure numerical stability. To evaluate the predictive readiness, a variety of unstructured mesh systems have been created, ranging from fine-resolution (5 m) configurations to coarse mesh (40 m) configurations. Forced scenarios, spatial averaged minute-by-minute AWS rainfall data were utilized using the Thyssen Polygon method to reflect the peak 6-hour intensity. The model performance was quantitatively evaluated by comparing the simulated maximum flooding range with historical flood footprint maps. The results indicate that the composite roughness approach reasonably preserves the hydrodynamic flooding characteristics at different mesh resolutions, including distance-aligned flow propagation. However, mesh coarsening has systematically increased the overestimated region, as the flooding classification is distributed among larger discrete computational units Analysis of the results of depth simulation shows that coarse numerical mesh tends to overestimate the extent of flooding, with shallower depths, while the distribution areas of deeper depths decrease. From an operational point of view, the analysis finds an important practical compromise: Mesh (10 m-20 m) at medium resolution can reduce the computational cost within the 10-minute update cycle, which is the standard for national flood forecasting, while maintaining essential spatial fidelity. On the other hand, overly rough meshes can risk inflating flood footprints, which can degrade predictive reliability.

ABSTRACT. Physically-based 1D-2D coupled models are essential for accurate flood analysis but are often limited by high computational demands, hindering their use in real-time forecasting. This study introduces KICT-UF-HPC, a novel high-performance framework designed to overcome these bottlenecks through heterogeneous computing. We implemented a hybrid parallelization strategy that allocates 2D surface flow calculations to GPUs via CUDA OpenACC and 1D sewer network computations to CPUs using OpenMP. The model's accuracy was rigorously validated against experimental flume data, while its efficiency was tested using the 2024 flood event in the Sangpyeong district, South Korea. The results show that our approach achieves a 35x speedup compared to conventional solvers while maintaining physical precision, demonstrating its potential as a robust core engine for real-time urban flood warning systems.

ABSTRACT. Accurate rainfall forecasting is a core requirement for anticipating hydrometeorological hazards—such as floods, typhoons, and droughts—whose frequency and intensity are increasing under climate change, and for reducing related damages. In particular, urban areas, characterized by high imperviousness and complex drainage systems, can experience a rapid escalation of inundation risk within approximately three hours after rainfall onset. However, numerical weather prediction and physics-based models often require substantial computation and processing time due to the complexity of atmospheric physical processes, which limits their practicality for real-time response in very-short-lead-time (0–3 h) rainfall forecasting. Accordingly, this study aims to perform very-short-term Quantitative Precipitation Estimation (QPE) and Forecasting (QPF) based on high-temporal-resolution radar rainfall data provided by the Korea Meteorological Administration (KMA), and to examine the applicability of the resulting rainfall products through linkage with urban inundation analysis. For QPE, an ensemble-based machine-learning algorithm, Random Forest (RF), is applied to calibrate radar-derived rainfall. For QPF, a NowcastNet-based very-short-term forecasting framework is employed, which separates and utilizes a rainfall motion field and a residual field to represent rainfall growth, decay, and advection. Finally, the applicability of the forecast rainfall is investigated by coupling the rainfall outputs with urban inundation analysis using SWMM (Storm Water Management Model) and the H12 (fully coupled 1D–2D) model. Overall, by linking very-short-term QPE and QPF with urban inundation analysis, this study is expected to complement the real-time limitations of conventional physics- and numerics-based approaches and to provide a foundation for decision-support systems for urban hydrometeorological disaster response.

ABSTRACT. Urban flood risk is increasing as extreme rainfall intensifies and storm patterns become more variable, challenging conventional flood protection and design standards. In densely developed river basins, intense precipitation interacting with complex infrastructure and drainage systems can produce rapid and highly localized inundation. The July 2018 Western Japan Heavy Rain caused catastrophic flooding and landslides, with severe impacts in Hiroshima, making the Ōta River Basin a compelling and well-documented case for high-resolution flood analysis. To move beyond traditional depth-only assessments, this study adopts a multi-metric flood hazard approach that jointly evaluates inundation depth, flow velocity, and flood duration. An XRAIN-driven unsteady 2D HEC-RAS model is implemented using the Diffusion Wave formulation, whose suitability at the basin scale is supported through validation against observed water levels. High-resolution XRAIN radar rainfall (250 m spatial and 1-minute temporal resolution) is used as the primary forcing, and its consistency with ground observations is evaluated using RMSE, bias and correlation. Model uncertainty is systematically assessed by testing four Digital Elevation Model (DEM) resolutions (1 m, 5 m, 12.5 m, and 30 m) and spatially varying Manning’s n values assigned based on land-use/land-cover classes. Preliminary validation against observed water levels shows that fine-to-intermediate DEM resolutions (1–5 m) consistently outperform coarse terrain representation, while the 30 m DEM produces large errors at multiple gauging stations (e.g., RMSE exceeding 4–5 m, compared with ~0.5–1.0 m for finer DEMs). Best model performance is generally achieved using 1 m or 5 m DEMs, although optimal resolution varies spatially across the basin. Roughness sensitivity analysis further indicates that land-use–based higher-resistance Manning’s n settings typically reduce simulation errors relative to minimum-resistance assumptions, highlighting strong combined controls of terrain resolution and hydraulic roughness on flood magnitude and persistence. Building on this uncertainty-aware hazard assessment, the validated framework will be applied to generate integrated depth–velocity–duration hazard layers and couple them with exposure and vulnerability indicators, enabling more robust identification of high-risk hotspots. Overall, this study delivers a reproducible, XRAIN-driven 2D modeling workflow that explicitly quantifies structural uncertainties and strengthens the reliability of hazard-to-risk translation for climate-resilient flood management in complex urban river basins.

ABSTRACT. The Kayu Ringin area in Bekasi City, West Java Province, located on the island of Java, Indonesia, is a low-lying urban catchment that is highly vulnerable to flooding. Flood events in this area are mainly caused by backwater effects from the Bekasi River into the Rawa Tembaga River during high-intensity rainfall, which inhibit effective gravity-based drainage. This condition is further exacerbated by rapid urban development and land-use changes that increase surface runoff while reducing infiltration capacity. Consequently, the existing drainage system is unable to operate adequately during extreme hydrological conditions, resulting in recurrent inundation.

This paper presents a technical justification for the implementation of a pumping-based flood control system along the Rawa Tembaga River, with a particular focus on the Kayu Ringin area. The primary objective is to establish a sound basis for determining design flood discharge, pump capacity, operational strategy, and intake structure configuration that are compatible with local hydraulic and topographic conditions. The study addresses key challenges related to flood discharge estimation, elevated river water levels, and the need for mechanical drainage solutions to mitigate flooding induced by backwater conditions in low-elevation urban areas.

The methodology comprises integrated hydrological and hydraulic analyses based on annual maximum rainfall data obtained from the nearest rainfall observation station. Design flood discharge is estimated using the Rational method, with the time of concentration determined according to catchment characteristics, including area, slope, and land cover. Flood hydrographs are developed to represent the temporal distribution of inflow, and flow routing analysis is performed to assess the interaction between inflow, available storage, and pump operation. On this basis, the required number of pump units, individual pump capacities, and operational water level thresholds are defined. In addition, the pump intake structure is designed in accordance with established hydraulic criteria to ensure stable inflow conditions and reliable pump performance.

The results show that the design flood discharge for the Kayu Ringin catchment, corresponding to a two-year return period, is 0.64 m³/s. The proposed flood control system consists of two submersible pumps, each with a capacity of 100 L/s, operated based on predefined water level thresholds. This configuration is capable of maintaining maximum water levels within acceptable limits, thereby reducing flood extent and duration. Overall, the findings demonstrate that the proposed system effectively enhances urban flood resilience and improves the reliability of drainage infrastructure in flood-prone urban environments, contributing to improved resilience against water-related hazards and disasters.

ABSTRACT. This study examines the water-surface rise and oblique jump formation in a tangential-inlet inclined chute, focusing on the theoretical relationships in related equations and the geometric configuration of oblique jump. The formation angle θs and contraction angle θc were measured using the X–Y projection method, and the downstream depth ratio h1/h0 was evaluated under varying inflow conditions. Results show that θs increases systematically with the approach-channel Froude number Fr0, confirming the dominant role of inflow inertia in shaping the oblique jump front. Comparisons with related equations reveal that classical horizontal-channel theory underestimates depth changes in inclined and tapered geometries, where additional gravitational and contraction-induced accelerations strengthen the jump. The findings provide a clear understanding of how geometric angles control oblique jump formation in inclined tangential inlets, offering useful implications for stabilizing inflow and improving deep-tunnel intake design.

ABSTRACT. Urban drainage networks constitute critical infrastructure for climate crisis response, and the quantitative assessment of their topological complexity and hydrologic response characteristics has emerged as an essential task for enhancing urban resilience. In particular, there is a rapidly growing demand for practical tools that enable systematic comparison and interpretation of drainage network topology across cities and quantitative evaluation of hydrologic sensitivity. Among the representative methods for quantifying the structural complexity of urban drainage networks, Gibbs’ model reconstructs river networks by probabilistically reassigning flow directions on a grid and concisely expresses the complexity of the entire network using a single parameter β. However, existing MATLAB-based implementations suffer from excessive computational loads during repeated β estimation and width function generation on high-resolution grids, imposing severe limitations on practical application and large-scale analysis. To address these challenges, this study fully reimplements the entire Gibbs’ Model algorithm in C++ and applies multi-level performance optimization techniques, including memory layout optimization, cache-friendly pointer operations, and compiler-level vectorization, to fundamentally eliminate computational bottlenecks. Furthermore, the optimized implementation will be extended into a high-efficiency tool that can be immediately deployed in real-world settings, supporting automated large-scale drainage network analysis pipelines, real-time coupling with hydrologic models, and quantitative evaluation of climate-adaptive urban resilience. Through these efforts, this research aims to overcome the computational performance limitations of the conventional Gibbs’ model and establish a technological foundation that enables real-time-scale topological analysis of city-wide urban drainage networks.

ABSTRACT. The bifurcation reach of the Asahikawa River, Okayama Prefecture, Japan is densely covered with riparian vegetation. During flood events, lodging and washout of this vegetation can occur and may significantly affect flow division between channels. Accurate prediction of such vegetation behavior is therefore essential for flood risk management. Conventional approaches often evaluate hydrodynamic forces on vegetation using simplified cylindrical drag models without distinguishing between stems and porous canopies, which may lead to inaccurate force estimation. In this study, two drag models that explicitly consider canopy porosity were compared and validated. In addition, a numerical framework for predicting lodging and washout of bamboo groves and woody vegetation was developed using drag moment and bed shear stress as evaluation indices. Flume experiments were conducted using resistance bodies that simulate tree canopies, and the effects of a porous media model and the Dupuit–Forchheimer (DF) drag model on water depth, velocity, and drag force were examined. Furthermore, three-dimensional flood flow simulations were performed for the July 2018 heavy rainfall event, incorporating vegetation distribution derived from airborne laser bathymetry (ALB) data. Based on these simulations, a judgment model using drag moment and dimensionless bed shear stress was applied to reproduce observed vegetation damage. The experimental results indicate that the model based on the DF drag model reproduces velocity distributions and drag forces with reasonable accuracy, whereas the porous media model tends to underestimate drag forces. Field-scale simulations demonstrate that the highest predictive accuracy is achieved when the DF drag model is used and drag moment alone is adopted as the evaluation index.In contrast, predictions based on bed shear stress are more sensitive to assumptions regarding sediment properties and incomplete vegetation data, leading to lower prediction accuracy. These results suggest that the DF drag model is effective for evaluating drag forces acting on porous vegetation canopies and that drag moment is a robust and practical indicator for predicting vegetation lodging and washout in three-dimensional flood flow analysis.

ABSTRACT. Recently, the frequency of extreme rainfall events, such as localized heavy rains, has been increasing due to the influence of climate change, leading to frequent flooding in urban areas. Consequently, damages to human life and property are on a continuous upward trend. To prevent and mitigate such urban flood damage, accurate field-based monitoring is essential. However, current monitoring of domestic sewer and drainage pipes primarily relies on water level measurements, calculating discharge through theoretical formulas. A significant issue arises during rainfall as various foreign substances on road surfaces enter the network and deposit in the pipes, forming diverse sediment layers. Therefore, the introduction of new monitoring techniques that can account for the effects of sedimentation is required. This study examines the variability of spectral information according to different media in monitoring the water depth of sewer and drainage pipes using spectral data. While depth monitoring utilizing spectral information has been studied for a long time, research targeting closed environments like sewer pipes, which lack specific light sources, remains insufficient. In such pipes, the presence of various debris and varying particle sizes affects reflection intensity, with wavelengths in the visible, blue, and green regions reacting sensitively to particle size. Thus, this study conducted indoor experiments to confirm the variability of spectral information based on sediment media typically found in actual pipes for accurate depth monitoring. A halogen light source was utilized with a luminous intensity of 720 lx. For precise results, experiments were conducted under darkroom conditions with various cases according to particle sizes and sediment heights. Spectral information was collected through a multispectral camera, and the collected images were used to estimate water depth through the Optimal Band Ratio Analysis (OBRA) technique. The results confirmed that the optimal band ratio and depth estimation accuracy varied depending on the particle size conditions. It was observed that the optimal band ratio combinations differed for each medium, suggesting that a combination effective for a specific medium may show degraded performance in others. This study evaluated and analyzed these factors through experimentation to improve the accuracy and applicability of pipe network depth monitoring using spectral information, which is expected to contribute to enhancing overall monitoring precision in sewer systems.

Acknowledgements This work was supported by Korea Environment Industry & Technology Institute(KEITI) through Climate Resilient R&D Project for Water-Related Disaster Management, funded by Korea Ministry of Climate, Energy and Environment(RS-2023-00218973)

ABSTRACT. Urban flood defense facilities are increasingly challenged by climate change, aging infrastructure, and the intensification of extreme rainfall events. While resilience has been widely discussed in urban flood management, most existing studies rely on static or index-based assessments at city or regional scales, offering limited insight into how individual facilities respond dynamically to disturbances. This study proposes a mathematical framework to evaluate facility-level resilience of urban flood defense systems and to identify critical thresholds that govern transitions between functional regimes. The framework integrates a composite sigmoid function to represent nonlinear performance trajectories of flood defense facilities under disturbance and recovery processes. Four resilience dimensions which are Robustness, Redundancy, Rapidity, and Resourcefulness are explicitly mapped to the parameters of the performance function, enabling initial structural performance, functional alternatives, recovery speed, and resource mobilization system over time. A comprehensive parametric analysis is conducted by systematically varying the four resilience dimensions across their feasible ranges, allowing exploration of diverse failure recovery pathways. Simulation results reveal the emergence of threshold behaviors and alternative regimes in facility performance. When resilience dimensions fall below critical levels, facilities are unable to recover to their pre-disturbance performance and instead stabilize at degraded functional states. In addition, scenarios involving repeated disturbances show that insufficient recovery between events can cause irreversible regime shifts, highlighting the importance of recovery timing under recurrent extreme rainfall. By linking resilience metrics directly to dynamic performance modeling, this framework advances understanding of multi stability and tipping points in engineered flood defense systems. The proposed approach provides a quantitative basis for identifying vulnerable resilience configurations and supports risk informed maintenance planning, adaptive investment prioritization, and proactive management strategies aimed at preventing undesirable regime shifts in urban flood defense facilities. Acknowledgement This work was supported by the National Research Foundation of Korea(NRF) grant funded by the Ministry of Science and Technology (RS-2024-00356786) and Korea Environmental Industry & Technology Institute grant funded by the Ministry of Environment (RS-2023-00218973).

ABSTRACT. Urban flooding has become increasingly frequent in urbanized areas due to the expansion of impervious surfaces and the increased occurrence of extreme rainfall events. To mitigate urban flooding, a Shinwol rainwater storage & drainage system has been constructed in the Yangcheon District of Seoul, Republic of Korea. However, the effectiveness of urban flood mitigation largely depends on the proper operation and efficiency of such infrastructure. Accordingly, this study focuses on the Shinwol stormwater storage facility as underground infrastructure and aims to develop a data driven water level prediction model to enhance its operational efficiency during rainfall events. The Shinwol stormwater storage facility is operated by controlling overflow gates based on water levels measured at an upstream point of a vertical inlet where multiple sewer conduits converge. Predicting the upstream water level is therefore critical for effective operation. However, physics based hydraulic models such as SWMM have limitations for real time operational use due to computational complexity, extensive parameter calibration, and scenario dependent uncertainty. To overcome these limitations, this study proposes an artificial intelligence–based water level prediction model using only observed sewer water level datas and rain datas. A deep neural network (DNN) is developed using observed sewer water level data from Yangcheon and Gangseo districts along with AWS rainfall data to construct a multivariate input vector representing the dynamic state of the sewer system. The DNN architecture consists of three hidden layers, enabling the model to capture nonlinear rainfall runoff and sewer response relationships. The target of the model is the predicted water level at the inflow junction point upstream of the vertical inlet which directly determines the inflow into the rainwater storage & drainage facility. The results demonstrate that the proposed model can stably predict sewer water levels without relying on any physics based hydraulic models. The proposed methodology enables observation based state prediction of underground shinwol stormwater storage facilities and highlights its potential as a core technology for smart urban flood control management.

Acknowledgements: This study was supported by Korea Environment Industry & Technology Institute(KEITI) through Technology development project to optimize This work is financially supported by Korea Ministry of Climate, Energy, Environment(MCEE) as 「Graduate School specialized in Climate Change」

ABSTRACT. Recent climate change has emerged as one of the most critical global challenges facing modern society, accompanied by a global increase in the frequency and intensity of extreme weather events. These changes have accelerated the occurrence of natural disasters such as floods, typhoons, and droughts, resulting in escalating economic and social impacts worldwide. In particular, urban flooding has become increasingly severe due to the combined effects of intensified heavy rainfall associated with climate change, the expansion of impervious surfaces driven by urbanization, and the limitations of aging urban drainage systems. Urban flooding extends beyond direct physical damage, disrupting critical infrastructure systems—including transportation, water supply, sewage, and power networks—and adversely affecting economic activities and social stability. As a result, it is widely recognized as a key factor that weakens overall urban resilience. Comprehensive assessment of urban flood risk therefore requires not only the evaluation of flood hazard characteristics, such as occurrence probability and inundation depth, but also the integration of exposure and vulnerability factors. Recent studies conceptualize urban flood risk as a function of hazard, exposure, and vulnerability, emphasizing the importance of structuring risk based on multiple interacting factors rather than relying on single indicators. Building on this perspective, the present study defines disaster impact factors as a comprehensive set of elements encompassing both exposure and vulnerability and proposes a framework for assessing urban flood risk at the urban scale. Ultimately, this study aims to support a more systematic and integrated evaluation of urban flood risk through the development of an urban flood risk matrix.

Acknowledgement This work was supported by the Technology Innovation Program (RS202400398858, Development of AI-based urban flood damage risk prediction and evaluation technology for practical use) funded By the Ministry of the Interior and Safety(MOIS, Korea)“

ABSTRACT. The frequency and severity of flood damage caused by climate change have been continuously increasing. Numerous previous studies on flood risk assessment have primarily focused on inundation area and damage magnitude. However, differences in regional recovery capacity and post-disaster recovery processes are difficult to incorporate into conventional risk assessment frameworks. Therefore, this study develops a quantitative method for assessing flood disaster resilience using the PEOPLES framework. The PEOPLES framework consists of seven domains: Population and demographics, Environmental and ecosystem, Organized governmental services, Physical infrastructure, Lifestyle and community competence, Economic development, and Social and cultural capital. The weights of sub-indicators within the PEOPLES framework were determined by applying the Fuzzy Analytic Hierarchy Process based on expert survey data, allowing uncertainty and subjectivity inherent in expert judgment to be systematically reflected. In addition, considering indicator redundancy and data limitations, a final set of core evaluation indicators was selected based on publicly available data. The proposed flood disaster resilience assessment method based on the PEOPLES framework captures not only direct recovery capacities, such as regional infrastructure and administrative response capability, but also indirect recovery capacities, including socioeconomic stability and community competence.

Acknowledgement This work was supported by the Technology Innovation Program (RS202400398858, Development of AI-based urban flood damage risk prediction and evaluation technology for practical use) funded By the Ministry of the Interior and Safety(MOIS, Korea)“

ABSTRACT. The increasing frequency and intensity of localized heavy rainfall caused by climate change have significantly increased the occurrence of urban flooding. In densely developed urban areas, intense rainfall over a short duration leads to rapid inflow into drainage networks, resulting in pressurization within conduits, backflow, surcharge, manhole cover displacement, and roadway inundation. These complex flood phenomena pose serious risks to urban safety and infrastructure. However, most existing urban drainage studies have relied primarily on numerical model–based analyses using 1D/2D models, while experimental validation data remain limited due to the difficulty of monitoring hydraulic conditions within underground drainage systems. As a result, the complex hydraulic behavior arising from the interactions among conduits, manholes, and downstream drainage facilities has not been sufficiently verified using measured data. To address these limitations, this study presents the development of a pilot-scale urban flood experimental facility designed to reproduce the integrated behavior of urban drainage systems under flood conditions. The facility is composed of a detention basin, stormwater inlets, conduits, manholes, downstream gates, and a storage tank, arranged to closely replicate real urban drainage configurations at near full scale. Rainfall inflow, discharge rates, and downstream boundary conditions can be precisely controlled, allowing a wide range of urban flood scenarios to be systematically simulated. The experimental facility is equipped with a high-resolution monitoring system capable of measuring discharge, water level, pressure, and structural responses in real time. These measurements enable quantitative evaluation of flood defense performance and resilience of individual drainage components as well as their combined system behavior. In particular, the facility allows direct observation of pressurized flow development, backflow, surcharge, and manhole-related failure mechanisms that are difficult to capture through field observations alone. The experimental data obtained from the pilot facility can be used to improve the reliability of urban flood modeling, support the calibration and validation of numerical models, and provide a scientific basis for evaluating flood defense capacity. Furthermore, the facility serves as a key experimental platform for improving urban drainage design standards and for developing and testing flood mitigation strategies based on Blue–Green–Grey infrastructure. Ultimately, this pilot experimental facility contributes to enhancing urban flood resilience through data-driven assessment and integrated flood management approaches.

Acknowledgements This work was supported by Korea Environment Industry & Technology Institute(KEITI) through Climate Resilient R&D Project for Water-Related Disaster Management, funded by Korea Ministry of Climate, Energy and Environment.(RS-2023-00218973)

ABSTRACT. Per- and polyfluoroalkyl substances (PFAS) and pharmaceuticals are persistent micropollutants that are poorly removed by conventional wastewater treatment due to their high stability and low adsorption affinity. In this study, a UV-regenerable iron-impregnated powdered activated carbon (Fe-PAC) system was developed and demonstrated at pilot scale (200 m³·day⁻¹) using real wastewater treatment plant effluent.

Fe-PAC functions as a dual adsorbent–catalyst, where contaminants are captured within carbon pores and subsequently degraded through UV/H₂O₂-induced photo-Fenton reactions at Fe active sites, enabling in situ regeneration of adsorption capacity. Without regeneration, short-chain PFAS (PFHxA, PFHpA) exhibited low and unstable removal, while long-chain PFAS (PFNA, PFOA) showed high initial removal followed by rapid performance deterioration due to site saturation. In contrast, periodic UV/H₂O₂ regeneration sustained PFAS removal above 95% after three months, with overall micropollutant removal efficiencies reaching up to 99.1%.

Pharmaceuticals such as diclofenac and florfenicol maintained consistently high removal (>90%) throughout long-term operation, and even hydrophilic caffeine showed recovery in removal efficiency after regeneration. No adverse effects on conventional water quality parameters were observed during continuous operation.

This pilot-scale study provides practical evidence that UV-regenerable Fe-PAC systems can overcome the intrinsic limitations of adsorption-based treatment and offer a robust, scalable solution for controlling persistent micropollutants in wastewater treatment facilities.

ABSTRACT. Recent studies introduced a hybrid of photovoltaic (PV) panels and green roofs to enhance panel performance, increase carbon sequestration, and mitigate urban heat. We (i) evaluated whether PV-green roofs are economic green urban infrastructures in a humid subtropical region and (ii) examined different panel spacings that optimize trade-off between PV-generated electricity versus carbon credit generated from green roofs. Traditional roofs with densely installed panels were the most economic roofing due to maximized energy generation and cooling from panel shading. Contrary to expectation, PV-green roofs were not the most economic due to devalued carbon credit and weak synergy from transpiration cooling enhancing panel performance. Annually, PV-green roofs produced 51850 kW with 3.3°C reduction in panel temperature and sequestered 2.5 tCO2 e and saved 122 kW heating/cooling energy via thermal insulation. Given the current carbon credit value, benefits from carbon sequestration covered only 2-4% of the net roofing cost. It required a 4-fold increase (up to 214 USD/ tCO2 e) in carbon credit to make PV-green roofs as economic as PV-concrete roofs. These results underscored a need to reconsider carbon credit value that will nudge stakeholders to take action towards a sustainable net-zero urban system that mitigates the impact of changing climate.

ABSTRACT. Green roofs provide multiple co-benefits—attenuating urban heat islandds, enhancing carbon uptake, and lowering building energy demand via superior thermal insulation—but their performance can be highly climate dependent. To assess their future effectiveness under warming temperatures and elevated atmospheric CO2, we applied MLCan, a multilayer canopy model, to compare vegetated and conventional roofs across projected climate scenarios. Model experiments indicate that insulation-driven buffering renders the green roof the environmentally and economically superior option in coming decades. In the peak summer month (July), green roofs lowered roof-level temperature by 4.3℃ (2041–2060) and 5.0℃ (2081–2100) relative to a traditional roof. Annually, the thermal buffering translates to energy savings of 17.4 kW m-2 and 21.2 kW m-2, corresponding to cost reductions of $94 m-2 and $114 m-2 in the near and far future, respectively. Carbon credits associated with CO2 uptake show no material differences among the four climate projections. When installation and maintenance, heating/cooling energy, and carbon credit revenues are combined, the green roof yields a 25% reduction in net cost—to $530 m-2—under the most perturbed climate compared with today’s climate. Overall, our results underscore that energy savings from enhanced thermal insulation are the dominant driver making green roofs a cost-effective, climate-resilient roofing strategy for mitigating heat stress and urban warming over the next several decades.

ABSTRACT. Urban wetlands play a critical role in regulating the urban water cycle and enhancing resilience against water-related stresses by providing ecosystem services such as water purification, ecological buffering, and carbon regulation. Despite their importance, conventional assessments of urban wetlands have largely relied on static, site-specific, and single-function approaches, limiting their effectiveness for proactive and sustainable urban water management.

This study develops and applies a Real-Time Data-Based Urban Wetland Integrated Management System designed to support continuous monitoring, integrated analysis, and decision-making for urban wetland management. The proposed system integrates natural and artificial water cycle data with ICT-based field monitoring information, including meteorological and environmental sensor data. A structured framework for data acquisition, quality control, database management, and visualization was implemented to ensure reliable long-term operation.

Ecosystem services are quantitatively evaluated across four key categories: water purification performance, biodiversity indicators, soil organic carbon, and carbon dynamics. These indicators are analyzed and visualized in real time, enabling the identification of temporal variations and functional interactions within urban wetlands. The system was applied to monitored urban wetland sites, demonstrating its capability to support integrated and data-driven evaluation of ecosystem service performance under varying conditions.

By enabling proactive management based on real-time data, the proposed system contributes to improving urban water cycle sustainability and strengthening resilience against water-related hazards. The framework is scalable and transferable to other urban environments, providing a practical tool for integrated urban water management, ecosystem service assessment, and policy support.

ABSTRACT. Accelerated urban expansion combined with climate change is amplifying extreme precipitation, heightening the risk that decentralized green infrastructure (GI) systems will experience imbalances between hydraulic load and storage capacity, ultimately leading to functional breakdown. However, prior research has largely concentrated on event-specific performance or short-term design considerations, offering limited insight into the threshold at which integrated grey–green drainage systems exceed their physical limits and transition into hydraulic failure under climate-induced extremes. This study develops an integrated evaluation framework to analyze system-wide hydraulic behavior and determine a climate-responsive design rainfall threshold. Design rainfalls were derived from CMIP6 shared socioeconomic pathway scenarios through bias-corrected projections. A green hydraulic stress index (GHSI) was formulated to quantify the spatial dynamics of hydraulic saturation across GI components, and a hydraulic safety margin (HSM) was further defined to convert flood mitigation performance into adaptation-oriented decision metrics. The results reveal that as design rainfall intensifies, risk distribution within the system undergoes marked spatial restructuring: areas classified as hydraulically stable decline progressively, while critical nodes expand, particularly under high-emission scenarios. A distinct tipping threshold is identified near 430 mm of design rainfall, beyond which system performance shifts from incremental deterioration to rapid and system-wide cascading failures. Moreover, HSM-based classification distinguishes three functional states—stable, transitional, and critical—thereby connecting GHSI diagnostics to practical adaptation strategies. These regimes clarify when routine management is adequate, when system optimization or retrofitting is warranted, and when structural expansion becomes essential, offering a robust foundation for long-term climate adaptation planning.

ABSTRACT. This work develops a numerical model based on the EPANET-DD (dynamic dispersion) model to simulate sodium chloride (NaCl) intrusion through a leak under negative pressure and the subsequent advective–diffusive–dispersive transport of the contaminant in the water network. Several modeling approaches have been developed in the literature to simulate the intrusion of contaminants from leaks. First, a model based on the classical orifice flow equation was used, in which the volume of contaminant intruded through the leak is assumed to be proportional to the square root of the internal-external pressure differential. This approach, adopted by LeChevallier et al. (2003), simplifies the phenomenon as it does not consider the effect of the porous medium surrounding the hole or the geometric variability of the leak hole. Its validity is limited to conservative estimates (worst-case) and ideal conditions, proving less accurate in complex real-world scenarios (Collins, et al., 2010). The model solves the solute transport equation under variable flow conditions using the random walk method, including axial and transverse dispersion processes in the pipelines and coupling the quasi-steady hydraulic simulation with the calculation of mass transport. Based on the study by Fontanazza et al. (2015), contaminant intrusion was simulated by imposing a transient subatmospheric pressure event (e.g., a water hammer) at a leak, causing the impulsive influx of NaCl into the pipeline. The model was calibrated and validated by comparing the simulations with experimental data obtained in the study by Fontanazza et al. (2015), showing an excellent ability to reproduce both the peak concentration and the propagation of the contaminant along the network.

ABSTRACT. Leakage remains a critical challenge for water distribution networks (WDNs). If not addressed promptly, it results in significant economic and environmental losses. While increasing the number of pressure sensors improves detection accuracy, financial constraints necessitate strategic sensor placement. Thus, determining optimal sensor locations is important. Previous research has utilized sensitivity matrices (SMs) to identify optimal locations by ranking their importance; however, these approaches often overlook detection overlapping, where multiple sensors monitor the same redundant areas. This study proposes a new SM-based methodology designed to maximize leak detectability by minimizing detection overlap. By leveraging the SM framework, the research identifies optimal sensor locations across various detection thresholds and leak magnitudes. The approach was tested on three distinct WDN types: distribution-oriented, US standard, and transmission-oriented networks. By utilizing SM, optimal configurations prioritize spreading sensors across the network to maximize unique detection coverage rather than clustering them. The results demonstrate that smaller WDNs achieve high detection rates with fewer sensors due to broader per-sensor coverage. In contrast, larger networks struggle to maintain high coverage even with more sensors, as individual sensor ranges are proportionally smaller. Ultimately, leak size and network topology significantly influence optimal sensor placement and overall detection capability.

ABSTRACT. Korean Water Management Master Plan projected that domestic water demand would increase by approximately 300 million m³ from 7.382 billion m³ in 2020 to 7.588 billion m³ in 2030, based on baseline demand. As domestic water demand increases, the likelihood of disputes related to inter-regional water allocation is growing. Particularly, since social, economic, and spatial factors are key determinants of domestic water demand, it is essential to analyze the spatio-temporal variability of domestic water demand based on these factors. This study utilized population as the social factor and GRDP as the economic factor, collecting domestic water demand data from 2000 to 2021. The regions of Busan, Ulsan, and Gyeongsangnam-do(province) were selected for this study, as they are the site of a proposed megacity development aimed at achieving balanced regional growth. Among the four indices for assessing spatial concentration, the Gini Coefficient (GC) was selected as the representative index through PCA analysis. A random forest model trained on data from 2000 to 2021 was used to forecast domestic water demand from 2025 to 2100. The results showed that in 2025, GC values for domestic water demand were highest in Busan, Ulsan, Changwon, Gimhae, and Yangsan, while in 2100, they were highest in Busan, Changwon, Ulsan, Gimhae, and Geoje. Temporally, GC values in the historical period gradually decreased, indicating a trend toward equilibrium. In the future, equilibrium is maintained initially; however, imbalance begins to intensify from 2087 under SSP1, from 2083 under SSP2, and from 2063 under SSP5. Spatially, clear distinctions emerged in regional balance or imbalance trends. This study visualized these results using Standard Deviation Ellipses (SDE) to identify where imbalances in domestic water demand are concentrated. These findings provide insights for regional balance planning and water resource infrastructure development in megacity regions.

ABSTRACT. River infrastructure plays an essential role in flood control, water utilization, environmental conservation, and public amenity, and is fundamental to reducing flood damage and supporting effective water resource management. Conventional river infrastructure management has primarily relied on periodic inspections and maintenance conducted on an individual-facility basis, with condition assessments and decisions regarding repair or reinforcement largely based on facility management ledgers. This approach represents a short-term and reactive strategy, in which functionally interconnected structures—such as levees, sluice gates, and drainage pump stations—are managed independently rather than as integrated units. As a result, inefficiencies arise in terms of management effectiveness and cost reduction. Classifying river infrastructure at the reach and district levels and implementing asset management can be regarded as an effective approach to improving management efficiency. Such asset management integrates related groups of facilities, including levees, sluice gates, and culverts, within river reaches or districts, and determines maintenance priorities through condition assessment and risk evaluation. Furthermore, this approach enables the reduction of long-term maintenance costs and allows for the integrated allocation of financial and human resources. In this study, conventional individual-facility-based maintenance practices are compared with integrated management approaches at the river reach and district levels from the perspective of river infrastructure management. Current domestic and international trends are reviewed through advanced asset management case studies, and future directions for improving river infrastructure management are proposed.

This study(work) was supported by Korea Environment Industry & Technology Institute(KEITI) through 'development of integrated asset management technology for water resources infrastructures to be incorporated in digital twins' (RS-2024-00337673), funded by Korea Ministry of Climate, Energy, Environment(MCEE).

ABSTRACT. In the context of accelerated transition toward integrated governance of water resources, water environment, and water ecology, alongside the synergistic advancement of pollution reduction and carbon mitigation, the deep integration of digital-intelligent (DI) technologies to reconfigure water quality system architecture and establish novel governance paradigm has become a critical initiative for addressing escalating water pollution challenges and fostering diversified water quality management frameworks. This study first establishes the core concepts and intrinsic characteristics of water quality systems based on time-domain functions and sequential data as foundational elements. Subsequently, leveraging DI technologies, a resilient five-layer hierarchical architecture for water quality systems is proposed, comprising "stereoscopic perception–model coupling–synergistic decision-making" as its key pillars. The regulation model underpinning this architecture is presented, alongside a detailed exposition of its core technological constituents. The innovative application of this technological framework, incorporating diverse intelligent algorithms across varied water quality monitoring and management scenarios, is then explored. Results indicate enhancements in operational efficiency exceeding 10% and a concomitant reduction in energy consumption of approximately 15% compared to conventional methodologies. Finally, potential challenges encountered during practical implementation of this system are analyzed, and viable future trajectories are outlined. This work aims to catalyze the profound integration of DI technologies with comprehensive water pollution management scenarios, thereby providing theoretical underpinning and actionable pathways for the intelligent advancement of water science.

ABSTRACT. A particle method for simulation of rain runoff in urban catchment has been developed and applied to a model terrain and a real urban terrain. Compared with the existing methods based on Shallow Water Equations, the proposedmethod is effective for reproducng the flow on the ground and around structures and buildings that dominate urban catchments. Inspite of the three dimensional computations only limited number of water particles need be solved, the computation is quite efficient. Althought the number of total particles are used to represent the overland flow tetermines the spatial resolution, for a relatively short rainfall over a limited urban terrain can be done quite efficiently with realistic results.

ABSTRACT. Malaysia is facing a rising threat from urban floods. The December 2021 Taman Sri Muda disaster not only displaced over 14,000 residents and claimed lives, but it also exposed notable preparedness and response gaps. This study determined changes in awareness, preparedness, social support, government perceptions, and non-government organisations (NGOs) assistance before and after the flood. Residents’ interest in future flood awareness programs and proposing strategies to improve disaster risk management at the community level were also established. In this study, 81 affected households participated in a structured survey, which enabled comparison of key indicators across pre- and post-flood conditions. According to the findings, substantial improvements were observed. Household planning and preparedness demonstrated the most significant increase, from 49.6% to 87.9%. Social cohesion and perceptions of institutional support also improved. Nonetheless, early warning delivery and program participation still exhibited weakness. The results also highlighted how local capacity was reshaped by lived disaster experiences, shifting risk perception into proactive preparedness and collective action. Consequently, this study developed a priority matrix integrating community support and agency feasibility to identify and prioritise high-impact, achievable flood preparedness initiatives. Furthermore, the data obtained offered evidence-based insights for agencies and policymakers centred on residents’ perspectives, which provided a roadmap to enhance the resilience of Malaysian communities in urban flood-prone areas and beyond.

ABSTRACT. Over the past few years, there have been numerous developments in the mountainous areas in the Philippines. Developing these areas involve alteration of the natural topography and cutting of trees, which lessens the ability of the soil to absorb the rainfall, increasing surface runoff, and contribute to the flooding, particularly in downstream areas. To address the growing concern of the flooding due to these developments, different stormwater management practices can be done and must consider the site characteristics, hydrological dynamics, and engineering design parameters. This paper focuses on incorporating a confluence basin into the existing waterways with downstream segment reconfiguration within a mixed-use development where two existing tributaries converge at a junction and the combined flow meanders downstream toward a coastal outfall. An in-depth study of the existing terrain was done to determine the optimized proposed realignment of the downstream segment to improve conveyance continuity and hydraulic grade control within the development corridor. In addition, a hydrologic analysis was conducted to establish the necessary considerations for hydraulic design of (i) a confluence pond that captures the combined tributary discharge, (ii) a road crossing via a two-barrel reinforced concrete box culvert, 2.75 m x 3.0 m per barrel, and (iii) the downstream realigned channel reach. The rubble masonry walls and the boulders ensure stability and help dissipate stormwater energy, which improves the hydraulic performance of the confluence basin and reconfigured downstream segment of the waterway by limiting local scour potential and managing high-velocity flow zones across transitions. This finding offers an effective approach to stormwater management that balances long-term resilience, functionality, and ecological compatibility.

ABSTRACT. Rapid urbanisation, climate change, and increasing water demand have placed considerable pressure on conventional water supply systems, particularly in developing countries. In Malaysia, the industrial and commercial sectors represent major consumers of treated water, intensifying concerns related to water security, supply reliability, and sustainable urban water management. Rainwater harvesting (RWH) has been recognised as a viable decentralised water management strategy that can reduce dependence on potable water supplies, mitigate surface runoff, and enhance resilience to climate variability. However, despite its potential benefits, the adoption of RWH systems in industrial and commercial sectors in Malaysia remains limited.

The purpose of this study is to identify the critical factors influencing the adoption of rainwater harvesting systems within industrial and commercial sectors in Malaysia. The research addresses key issues related to organisational decision-making, technical feasibility, economic considerations, and regulatory support that affect the implementation of RWH systems in non-residential buildings. A quantitative research approach was carried out, employing a structured questionnaire survey targeting facility managers, building owners, and decision-makers from industrial and commercial establishments across selected regions in Malaysia. The collected data were analysed using descriptive statistical techniques and multivariate analysis, including Structural Equation Modelling (SEM), to examine the relationships between influencing factors and RWH adoption intentions.

This study identifies the most significant drivers and barriers affecting RWH adoption, such as initial investment costs, perceived long-term economic benefits, technical reliability, organisational awareness, and the adequacy of existing regulatory frameworks. The findings provide empirical insights that support policymakers, urban planners, and industry stakeholders in developing targeted strategies, guidelines, and incentives to promote the wider implementation of RWH systems. Overall, this research contributes to climate change adaptation and sustainable urban water management by strengthening the role of rainwater harvesting in enhancing water resource resilience in industrial and commercial contexts.

ABSTRACT. Flooding has long been considered the final outcome of extreme atmospheric forcing. Flood events, however, have the role of land surface features that can actually affect the evolution of the atmosphere. In particular, the flood-induced changes in the coverage of surface waters have significant impacts on the land surface characteristics, which in turn have considerable control over hydro-atmospheric coupling. In this regard, the flooding impacts the development of the boundary layer and the thermodynamic evolution of the atmosphere, which are not sufficiently studied in hydrology and hydraulic engineering. In this context, the main problem is the determination of the spatial and temporal distribution of flooding and its relation to the responses in the atmosphere. In the present study, the effect of flooding on the evolution of Convective Available Potential Energy (CAPE) is studied based on hydro-atmospheric coupling, considering the AI-driven flood extent areas using Synthetic Aperture Radar (SAR) data. The analysis reveals the relationship of the SAR-based flood-induced changes in surface water coverage with variations in surface energy fluxes, development of the boundary layer, and atmospheric instability. The results indicate that an increase in surface water coverage as a result of flooding is strongly related to an increase in hydro-atmospheric coupling, mainly in the form of moisture recycling from the land surface to the boundary layer. An increase in the extent of inundated surface areas is related to an increase in latent heat flux, which in turn increases near-surface moisture availability and CAPE. This research suggests that as the amount of flooding detected by SAR observations increases, the sensible heat flux is significantly reduced due to the suppression of surface heating. Such a suppression affects the growth of the boundary layer and vertical mixing, causing a reduction in atmospheric instability and hence a reduction in CAPE despite the presence of elevated moisture within the lower layer of the atmosphere. An analysis of the two cases discussed in the research demonstrates that the effect of flooding on CAPE is nonlinear. Overall, this study contributes to the understanding of the dynamic modulation of boundary layer development and convective potential through hydro-atmospheric coupling processes. The explicit inclusion of flood extent information from SAR data in the results also emphasizes the importance of the dynamically observed state of surface waters in understanding extreme precipitation processes and in flood risk management in increasingly extreme hydro-meteorological situations.

ABSTRACT. on infrastructure as a result of climate change, poor urban planning, and the rapid expansion of impervious surfaces. These challenges highlight the need for small-scale, decentralized nature-based solutions (NbS), particularly low-impact development (LID) technologies, which can enhance local climate adaptation and urban resilience. This study assessed the performance of three LID practices surface constructed wetlands (SCW), rain gardens (RG), and tree box filters (TBF) with respect to stormwater pollutant removal, mitigation of the urban heat island (UHI) effect, and soil organic carbon (SOC) sequestration. The findings revealed that impervious areas such as roads and parking lots were major contributors to urban stormwater pollution. Among the evaluated LID systems, SCWs demonstrated the highest pollutant removal efficiencies, achieving reductions of up to 67% for total suspended solids (TSS), 69% for total nitrogen (TN), and 81% for total phosphorus (TP). This superior performance was mainly attributed to effective sedimentation processes, well-designed filter media, and the presence of mature vegetation. In terms of thermal regulation, all LID technologies showed significantly lower surface temperatures up to 7 °C cooler compared to surrounding impervious surfaces, emphasizing their strong potential for reducing UHI effects.SOC stocks varied between 26.3–70.3 Mg/ha at inflow sections and 33.3–80.2 Mg/ha at outflow sections, with SCWs recording the highest values. Higher SOC accumulation was associated with vegetation root biomass, system maturity, and the deposition of organic matter transported by runoff. The 18% increase in SOC at outflow areas indicates progressive carbon accumulation over time, particularly in sediment-rich zones, suggesting long-term carbon storage potential and reduced greenhouse gas re-emissions. Overall, the study underscores the role of LID technologies as sustainable, multifunctional strategies for resilient urban development under escalating environmental pressures.

ABSTRACT. Dry weather flows (DWF) are non-stormwater runoffs from urban stormwater drainage systems that can jeopardize receiving water quality. DWFs often originate from illicit connections and inappropriate discharges. While high-frequency sensor networks have expanded data availability, effectively and accurately interpreting these datasets at the sub-catchment scale remains a challenge. Currently, a significant research gap exists regarding the use of long-term, multi-parameter data to quantify DWF pollution concentration trends and discharge patterns. Furthermore, few studies have attempted to correlate these high-resolution observations with specific source types caused by human activities. Consequently, this study aims to systematically quantify baseline behavior, the variability of discharge events, and the temporal signatures of degraded water quality as distinct responses to human activity. The study employs a combined observational and analytical methodology utilizing high-resolution sensor data. A robust framework for DWF pulse detection was developed to identify discharge events from background noise based on slope and amplitude thresholds. Subsequently, these pulses were analyzed to identify recurring temporal patterns across diurnal and weekly cycles. Clustering methods were then utilized to investigate common patterns in the observed water quality states at different time periods. The analyses are used to investigate whether DWF timing exhibits repeatable patterns within the catchment and to what extent such patterns might be consistent with anthropogenic scheduling. Heatmap-based visualizations are applied to explore the recurrence of event timing, and clustering outputs are used to describe water quality states and potential transitions between baseline conditions and pulse events. In addition, the workflow examines the duration of degraded conditions following disturbances and the apparent recovery behavior of the system. By integrating high-frequency monitoring with structured event-based analysis, this research provides an innovative diagnosis framework for DWF within urban drainage subsystems. The identification of common water quality states and timing signatures offers a promising solution for DWF source tracking. Ultimately, this approach helps to extract robust insights into DWF features while acknowledging the inherent limitations of field sensing.

ABSTRACT. This study focuses on an important issue that has a substantial impact on the Filipino people's well-being. The study looks at the link between water quality, public health, and sanitation practices in the Philippines, which faces several issues due to waterborne infections and poor sanitation infrastructure. The study technique combines quantitative and qualitative analyses to thoroughly explore the extent and severity of waterborne illnesses common in the Philippines. It identifies the individual infections and pollutants that cause outbreaks, as well as the implications for public health. Furthermore, the study assesses the efficacy of present sanitation. strategies and their effect on lowering the incidence of Waterborne illnesses. The study attempts to uncover public health risks caused by poor water quality and inadequate sanitation. It also catalyzes change, inspiring action to address the issues of waterborne illnesses in the Philippines. This study intends to improve public health and quality of life for countless Filipinos by assuring access to safe and clean water sources, as well as strong sanitation practices. This study aims to improve our understanding of the link between the state of water, public health, and sanitation. It gives a forum for emphasizing the significance of raising water quality standards and implementing better sanitation methods. The findings are likely to inform policy suggestions and public health actions, resulting in a better and more sustainable future for the Filipino people.

ABSTRACT. Super-resolution (SR) offers a pathway to bridge the spatial-detail gap between freely available Sentinel-2 imagery (10 m) and commercial very-high-resolution sensors such as PlanetScope (~3 m), yet SR benefits are often evaluated only by image-fidelity metrics rather than by improvements in downstream environmental mapping. This study develops a convolutional neural network (CNN)-based SR framework to enhance 10 m Sentinel-2 reflectance to an effective 3 m product using PlanetScope imagery as the supervisory target, and then tests whether the SR Sentinel-2 data improve river morphology classification relative to the original 10 m Sentinel-2 inputs. Sentinel-2 and PlanetScope scenes were paired patches were constructed for supervised training of the SR CNN. SR performance was quantified using standard fidelity measures (e.g., PSNR and SSIM) and visual inspections focused on river-relevant structures such as channel edges, sediment bars, and narrow riparian strips. For the downstream task, we trained a Random Forest (RF) classifier to map four river corridor classes—water, sediment, tree, and grass—using (i) the original 10 m Sentinel-2 imagery and (ii) the 3 m SR Sentinel-2 imagery, with identical training samples, feature sets, and validation design to ensure a fair comparison. Classification skill was assessed using macro-averaged F1-score and class-specific precision/recall, supplemented by confusion matrices and spatial error patterns across representative reaches. Results show that the CNN SR model reconstructs sharper and more continuous riverine features than conventional upsampling and improves the separability of spectrally similar classes at sub-reach scales. Compared with the 10 m baseline, RF classification using SR Sentinel-2 yields higher accuracy and more coherent maps of waterlines, exposed sediment, and narrow vegetation belts, demonstrating that SR can deliver practical gains for river morphology mapping while retaining Sentinel-2’s wide coverage and revisit frequency. The proposed SR-to-classification workflow provides a transferable approach for enhancing river corridor monitoring in data-sparse regions and supports near-operational mapping where PlanetScope coverage is limited or cost-prohibitive.

Funding: This work is financially supported by Korea Ministry of Climate, Energy, Environment (MCEE) as Research and Development on the Technology for Securing the Water Resources Stability in Response to Future Change (RS-2024-00332494).

ABSTRACT. Image-based river monitoring has gained increasing attention in flood detection and river environment analysis. This trend is driven by the growing availability of satellite imagery, unmanned aerial vehicles (UAVs), and closed-circuit television (CCTV) systems. However, little effort has been devoted to systematically constructing and publicly sharing benchmark river image datasets for rivers in Korea. In addition, the influence of environmental observation conditions, particularly illumination variability, on feature-based image analysis methods has not been sufficiently examined using multi-site observations. These limitations hinder the reliability and interpretability of image-based river monitoring techniques in practical applications. The objective of this study is to quantitatively investigate the effects of illumination conditions on feature-based river image analysis using a publicly available dataset that reflects spatiotemporal variability across major river basins in Korea. The analysis is conducted using the KU River Image Dataset, which consists of 1,660 river images collected under diverse observational conditions and includes metadata describing acquisition time and location. To generalize illumination effects beyond a single site, time-series CCTV images acquired from multiple observation locations within the Yeongsan River basin are analyzed. Feature detection and image matching are conducted using feature-based methods, including the Scale-Invariant Feature Transform (SIFT). The results reveal clear and consistent differences in keypoint detection characteristics between daytime and nighttime conditions across observation sites. The octave at which the number of detected keypoints and the mean matching rate are maximized differs markedly depending on illumination conditions. For image pairs acquired under mixed illumination conditions, matching rate is substantially reduced. Low average matching rates are observed across all octaves and sites. These findings suggest that illumination variability plays an important role in shaping feature-based image matching results, beyond local site-specific factors. This study provides multi-site evidence that illumination conditions constitute a critical environmental factor in feature-based river image analysis. The findings highlight the necessity of explicitly accounting for illumination variability when developing, evaluating, and applying image-based river monitoring methods, and contribute to improving the robustness and practical applicability of vision-based flood monitoring systems.

This work was supported by Korea Environment Industry & Technology Institute(KEITI) through R&D program for innovative flood protection technologies against climate crisis, funded by Korea Ministry of Climate, Energy and Environment(MCEE)(RS-2023-00218873)

ABSTRACT. Hydrodynamic modeling provides insights into the physical state of a water body by simulating processes based on established physical equations, conservation laws, and boundary conditions. However, errors from input data and model structure cause divergence in model behavior, which eventually degrades prediction quality. Data assimilation helps mitigate model divergence by systematically incorporating available observations into the initial predicted state; however, the limited spatiotemporal availability of observations poses challenges to its effective implementation. This study implemented Rank Histogram Filter to assimilate temperature measurements in the Ogouchi Reservoir, and the resulting updates were subsequently interpolated onto the model grid using kriging. This framework preserves the intrinsic statistical relationships between model cells and ensures that the spatial correlations and other system patterns are maintained consistently throughout the interpolation process. Moreover, correlation matrices were modified using a distance-based tapering function to limit the effects of spurious correlations between physically distant points. Predictions from twenty ensembles were assimilated using a fixed point (FM) and multiple longitudinal measurements (MLM), and the results were compared against unassimilated baseline models (Base). Evaluation based on Root Mean Square Error statistics and updated vertical temperature profiles showed substantial improvements in forecast qualities of assimilated states. The assimilation using MLM achieved the best performance as states from multiple localities were represented instead of relying on a single observation. Examination of horizontal temperature variations revealed that kriging interpolation can generate smooth, continuous spatial fields while preserving the reservoir’s underlying hydrodynamic patterns. The combined framework also captured local phenomena that would otherwise be lost during model progression, underlining its ability to retain small-scale thermal heterogeneity. The improvements achieved through the integrated process were found to dissipate as the updated models were propagated over a five-week period, indicating the importance of selecting optimal assimilation intervals. The results demonstrate the potential of combining data assimilation and geospatial interpolation techniques to improve understanding of thermal processes, support reservoir monitoring and management, and address challenges associated with sparse observations.

ABSTRACT. The settling and transport of inertial particles in coastal environments matter for many environmental issues, including transport and dispersion of sediment, microplastic and other debris. Because inertial particles do not simply follow the carrier-fluid displacement, their behavior can be strongly nonlinear and is expected to depend on both wave conditions and particle properties. In this study, we perform laboratory experiments with negatively buoyant spherical particles with specific gravity from 1.3 to 2.5 in regular surface gravity waves and investigate phase-resolved slip dynamics. Particle velocities are measured using particle tracking velocimetry (PTV), and the surrounding Eulerian wave orbital velocity is obtained simultaneously from particle image velocimetry (PIV). By combining these measurements, we estimate the instantaneous slip velocity at the particle position and compute wave phase using Hilbert transform. Experimental results were analyzed with phase-ensemble statistics of both horizontal and vertical slip components. They show clear phase-locked modulations of slip, and modulation amplitudes and phase shifts from wave phase change systematically with particle inertia and wave conditions. To interpret these trends, we also derive an asymptotic solution for the slip velocity by solving the particle equation of motion (Maxey-Riley equation) with a nonlinear drag model. Using a perturbation expansion with respect to wave nonlinearity, the solutions are expressed as periodic functions of wave phase and predict first- and second-harmonic contributions to slip modulation. We then compare phase-averaged measurements with the asymptotic predictions and find consistent trends in both modulation magnitude and phase shift across cases. In particular, cases with larger particle Reynolds number exhibit stronger slip modulation and larger phase lag. The enhanced relative motion tends to occur around phases when the vertical wave acceleration changes their directions, indicating that wave-driven forcing can locally amplify settling during specific phase of the wave-cycle. These findings support the concept of preferential settling under waves and provide combined experimental-theoretical framework for incorporating phase-dependent particle inertia effects into coastal transport models.

ABSTRACT. The objective of this study is to propose a cluster-based water quality estimation framework that integrates physical and virtual sensors while explicitly accounting for spatial heterogeneity and nonlinear water quality responses in large river basins. Conventional single-model approaches often fail to capture spatially diverse water quality dynamics driven by heterogeneous watershed characteristics, pollutant sources, and hydrological conditions. To address this limitation, monitoring stations with similar water quality characteristics were grouped into clusters, and independent prediction models were developed for each cluster to improve estimation accuracy and applicability. The proposed framework was applied to the Han River Basin in South Korea, including the main stem of the Han River as well as the Namhan and Bukhan sub-basins. Long-term observations from the national water quality monitoring network spanning 2016–2025 were utilized. Sensor-based water quality variables, including water temperature, pH, dissolved oxygen, electrical conductivity, chlorophyll-a, suspended solids, nitrate nitrogen, ammonia nitrogen, and orthophosphate, were used as input features. Spatial clustering was conducted using K-Means, hierarchical clustering, and Gaussian Mixture Models based on station-level mean water quality characteristics. The optimal cluster structure was determined using silhouette coefficients and complementary performance metrics. For each cluster, regression models based on Random Forest, XGBoost, and multilayer perceptron (MLP) algorithms were developed. Model performance was evaluated through cross-validation and hyperparameter optimization. The results demonstrate that the cluster-based approach consistently outperformed single-model frameworks for both T-N and T-P estimation, with particularly strong predictive capability for T-N concentrations. In contrast, T-P estimation exhibited pronounced performance variability across clusters, reflecting the nonlinear behavior and higher prediction difficulty associated with low-concentration and highly variable phosphorus dynamics. Incorporating orthophosphate as a direct phosphorus-related indicator within the hybrid sensor framework further improved T-P estimation performance. Overall, the proposed framework reduces spatial uncertainty in water quality prediction and provides a robust methodological basis for adaptive water quality management in large river basins under increasing climatic and hydrological variability.

This work was supported by the Ministry of Climate, Energy, Environment through the project 'Research and Development on Technology for Securing Water Resources Stability in Response to Future Changes (grant number RS-2024-00332114)'.

ABSTRACT. Purpose of the work: The purpose of this study is to validate the accuracy of a numerical model by directly comparing numerical results with detailed field observations in the Kurobe River fan, a steep mountain river. Specifically, the study applies a calculation method based on Riemann invariants, which uses only water levels as boundary conditions without requiring discharge data. The research quantitatively evaluates how these results align with observed data, including spatial flow velocity and water surface elevation obtained through UAV-PIV and UAV-LiDAR, as well as discharge measurements using ADCP. Key issues: Verification data for flood flows in natural rivers has traditionally been limited to point-based measurements using water level gauges, making it difficult to sufficiently verify the reproducibility of spatial flow patterns. Furthermore, while comparison with actual phenomena is essential to ensure the reliability of numerical calculations, there remains the challenge that calculation results are highly dependent on the specified boundary conditions and parameters, particularly discharge and riverbed roughness coefficients. This dependency makes it difficult to purely evaluate the validity of the model itself or the reproducibility of the physical phenomena. Methodology: The numerical model "Cabernet2D," an iRIC solver, was used for the analysis. A boundary condition treatment based on Riemann invariants was implemented at the upstream and downstream ends, where measured water levels were provided as the only known quantities, and discharge was treated as an unknown variable. For validation, surface flow velocity distributions via UAV-PIV, high-density water surface elevation data via UAV-mounted LiDAR, and discharge data via ADCP were collected and directly compared with the numerical results. Results and conclusion: The results demonstrated that the flow distribution and conditions in each channel were reasonably reproduced, even though discharge was not provided as a boundary condition. The calculated water levels successfully captured the transitions between subcritical and supercritical flows occurring near groynes and steep sections, showing agreement with UAV-LiDAR measurements within an order of tens of centimeters. This study demonstrated that the results of the analysis using only water level information can be quantitatively evaluated through detailed spatial data from non-contact measurements. This provides a practical framework for objectively verifying the validity of numerical simulations in natural river analysis, where boundary conditions are often constrained.

ABSTRACT. High concentrations of PM2.5 in the ambient atmosphere of the Indian Subcontinent has been evident over several decades causing a health burden to citizens and environmental degradation especially in the metropolitan cities. Spreading awareness through deeper research is crucial for planning preventive actions and enforcing strict regulations towards building a sustainable and healthy environment. This research aims to bridge this gap using CMIP6 General Circulation Models, satellite-derived dataset and a soft computational technique, to predict the pollutant concentration under influential parameters and across multiple scenarios. The input features selected for this process were the CMIP6 GCMs including the criteria pollutant PM2.5 and four significant meteorological parameters while Satellite-derived PM2.5 was the dependent variable. These variables were extracted at the latitude-longitude coordinates of GCM grid points adjacent to the target city. The Support Vector Regression technique modelled and validated the historical data with satellite values from 1998 to 2014 with acceptable results, thereby displaying confidence in predicting the pollutant concentration from 2015 till 2100 across ssp245 and ssp585 scenarios. The SVR derived MIROC6 to be the most compatible GCM for Pune and Mumbai with 0.92 ≤ R ≤ 0.96, RMSE ≈ 7 µg/m3 and MAE ≈ 5.5 µg/m3 while Delhi showed considerable variations in GCM responses as well as the respective error measures. However, since satellite data is collected remotely, model validation with ground observations was found essential and the average result was 0.71 ≤ R ≤ 0.81, 11 ≤ RMSE ≤ 51 and 9 ≤ MAE ≤ 40 for the three cities. The 10-year average results indicate an increase in the PM2.5 concentrations till 2044 and the following period from 2045 till 2100 show a 43%, 6% and 36% decline over Delhi Pune and Mumbai for ssp245, while the decline in the ssp585 scenario was found lesser as 26%, 22% and 18% for the three cities respectively. In the decade spanning 2015-2024, the 6 monthly average pollutant concentration from Oct through Mar was nearly 150, 50 and 55 for Delhi, Pune and Mumbai respectively. The long-term predictions obtained by integrating simulations and soft-computing in light of influential parameters are presented in this research for planning suitable mitigations strategies to diminish the ill-effects PM2.5 on the environment and its inhabitants.

ABSTRACT. Hong Kong experiences some of the world’s most intense rainstorms, with individual extreme events causing over US$100 million in damages and repeatedly overwhelming low-lying districts such as Happy Valley. Although the Happy Valley Underground Stormwater Storage Scheme (HVUSSS) was implemented in 2015 to mitigate flash flooding through smart controlled movable weirs and SCADA based real time operation, its rule-based approach lacks predictive capability, raising concerns about its resilience as rainfall becomes increasingly severe in the face of a changing climate. These concerns highlight the need to review HVUSSS performance using data-driven analysis rather than its operating rules. This study addresses the gap by i) integrating Peak Over Threshold (PoT) analysis, ii) climate adjusted return level projections, and iii) one dimensional (1D) hydraulic model stress testing, all performed using an AI powered Python workflow that automates data processing and ensures reproducibility. Daily rainfall records from Hong Kong Observatory stations (2015–2025) are analyzed using PoT thresholds to characterize precipitation extremes; the resulting exceedances are aligned the Intergovernmental Panel on Climate Change (IPCC) Sixth Assessment Report (AR6), to derive climate aware 50-, 100-, and 200-year return levels. These projected extremes are applied to a 1D hydraulic model simulating HVUSSS to quantify impacts on peak water levels, effective storage requirements, drawdown times, internal flow distribution, and spill frequencies under extreme scenarios. Preliminary results with 95th percentile rule from PoT indicate that Happy Valley now experiences an extremely wet month approximately once every 19 months, and the projected return level is approximately 2.1 (50-year), 2.7 (100-year) and 3.4 times (200-year) more than the maximum daily rainfall record observed. Together, these analyses generate climate informed insights that directly enhance HVUSSS performance. Outputs include updated climate responsive design curves for the Happy Valley catchment, refined hydraulic performance metrics, and actionable operational recommendations, such as preemptive drawdown triggers linked to short range extreme rainfall forecasts, dynamically adaptive weir control, and contingency allowances for partial storage availability or temporary blockages. By coupling PoT based rainfall characterization with climate aware return level estimation and operational stress testing, the study provides a transparent, auditable, and transferable framework that surpasses the existing reactive SCADA rule based approach, demonstrating how predictive, risk informed strategies can strengthen storage reliability, improve peak flow attenuation, and reduce flood risk while offering a practical blueprint for similarly dense coastal basins facing rapidly intensifying rainfall extremes.

ABSTRACT. Cox’s Bazar in Bangladesh is the most dynamic coastal regions, experiences persistent erosion and accretion that threaten infrastructure, ecosystems, tourism, and livelihoods. Understanding short- and long-term shoreline dynamics and their physical drivers is essential for effective coastal management. However, many previous assessments are constrained by sparse temporal sampling and limited integration between shoreline observations and hydrodynamic forcing. The purpose of this study is to quantify multi-decadal shoreline evolution and to examine how seasonal and interannual wave forcing governs spatial and temporal shoreline variability along the Cox’s Bazar–Teknaf coast during 1988-2024. Shoreline positions were extracted from multi-mission Landsat and Sentinel satellite imagery using the CoastSat toolkit, an automated, machine-learning-based shoreline detection tool with dense temporal coverage. Shoreline trends were quantified using End Point Rate (EPR), Linear Regression Rate (LRR), and Error-Weighted Regression (ERR), and Empirical Orthogonal Function (EOF) analysis was applied to identify dominant modes of shoreline variability. Nearshore wave conditions were simulated with the Simulated WAve Nearshore (SWAN) model to estimate significant wave height, swell, wave period, and wave directional spectrum. All extracted shoreline positions were tidally corrected to reduce water-level bias and improve trend reliability. Results show strong spatial and temporal heterogeneity along the 90 km littoral cell. Several segments exhibit long-term erosion rates up to 13.0 m/year to episodic accretion up to 17.0 m/year. EOF decomposition indicates coherent large-scale sediment redistribution together with localized phases of erosion (1988-2000), recovery (2001-2010), and stabilization (2010-2024). SWAN simulations highlight a dominant monsoon-driven south-westerly wave regime, with higher wave exposure corresponding to multiple erosion-prone zones. Shoreline response to wave forcing is spatially heterogeneous and not strictly proportional to modelled wave energy. While elevated wave exposure often coincides with enhanced erosion, persistent retreat also occurs where mean wave heights are lower, indicating that wave forcing provides the energetic context but not the sole driver while sediment availability and alongshore redistribution, human intervention govern the shoreline dynamics. Wave forcing also exhibits a lagged influence on shoreline adjustment, with erosion and recovery occurring weeks to months after peak wave energy due to delayed sediment redistribution. The novelty of this work lies in coupling high-frequency satellite-derived shorelines with process-based wave modelling to resolve long-term trends and lagged shoreline responses that are not captured in conventional snapshot-based studies. Ongoing and future analyses will further explore process coupling, human intervention, and predictive shoreline response under evolving wave climates, supporting improved coastal protection and climate-resilient shoreline management in Bangladesh.

ABSTRACT. Coastal environments are among the most valuable and dynamic regions on Earth, supporting diverse ecosystems, high biological productivity, and socioeconomic activities. These coastal regions are increasingly exposed to hazards such as coastal flooding and erosion. The risks caused by these hazards depend not only on the level of hazards but also on the vulnerability of the exposed coastal environments. Therefore, it is important to have a robust understanding of the coastal vulnerability. The Coastal Vulnerability Index (CVI) is the most widely used approach for assessing coastal vulnerability, integrating multiple indicators into a single relative measure of vulnerability. While CVI assessments have been applied at local and regional scales, a consistent global scale coastal vulnerability assessment, which is important for informing the macro-scale decisions, has not been attempted to date. Here, by employing currently available global datasets together with the CVI approach, we assess present-day global coastal vulnerability, and analyze the contribution of each indicator to the CVI to understand the dominant indicators. The CVI at the global coast were calculated by four selected commonly used CVI methods in the literature. These CVI methods aggregate multiple geophysical and coastal forcing indicators (including: coastal slope/relief, geology, geomorphology, wave height, relative sea level change rate, mean tidal range, and shoreline change rate) to determine the vulnerability level of coastal environments. Openly available global coastal datasets were used to extract the data for each indicator at 363,424 sampling locations over the global coastline. Globally computed CVIs from the four methods were compared with higher resolution local and regional CVI assessments to identify the optimal method for a global scale assessment. The CVI method thus identified was then adopted to assess present-day coastal vulnerability (from ‘Very Low’ to ‘Very High’ level) at country and IPCC AR6 region level globally. The CVI country level assessment shows that Aruba, Benin, Togo, Democratic Republic of the Congo, Bonaire, Sint Eustatius and Saba, Sri Lanka, Nigeria, French Guiana, Ghana and Liberia have the highest median coastal vulnerability at present. At the IPCC AR6 region scale, Central North America, and Northern South America emerge as the regions with the highest median coastal vulnerability. Both country level and IPCC region level results indicate that tropical and subtropical regions are more vulnerable to coastal hazards. In countries with ‘High’ to ‘Very High’ median CVIs, the dominant contributors to the present-day coastal vulnerability are geomorphology, mean tidal range, and coastal slope.

ABSTRACT. Accurately predicting the transformation of random waves in the nearshore zone requires numerical models capable of representing both nonlinear wave evolution and the depth-induced energy dissipation associated with wave breaking. Although frequency-domain mild-slope formulations effectively describe shoaling and refraction, their accuracy typically decreases as waves approach breaking, particularly when higher-order statistics, spectral distortion, and nonlinear interactions become significant. To address these limitations, a nonlinear frequency-domain wave model is formulated to more consistently account for nonlinear interactions, horizontal depth variation, and modulation scales, while a modified frequency-dependent dissipation function is introduced to better represent breaking-induced energy dissipation across the spectrum. In this work, the two components are combined to simulate unidirectional random waves propagating over a sloping beach, and the model performance is evaluated against laboratory measurements of free-surface elevation at multiple nearshore locations. Results show that applying dissipation uniformly across all frequencies leads to numerical instability and reduced accuracy in the shallowest regions, whereas incorporating a frequency-weighted dissipation term produces stable solutions across the domain and substantially improves predictions of spectral evolution, energy decay, and higher-order wave characteristics. The enhanced model successfully captures the progressive distortion of wave shape, the attenuation of higher-frequency components, and the increasing influence of nonlinearity as waves approach breaking. Overall, the integrated modeling framework provides a more robust and physically consistent tool for simulating nearshore random wave transformation, demonstrating clear advantages in conditions where nonlinear effects and breaking dominate the wave dynamics. These findings underscore the importance of incorporating meaningful frequency dependence into breaking-induced dissipation and highlight the potential of the improved model for applications in coastal engineering design, wave forecasting, and hazard assessment.

ABSTRACT. Nature-based shore protection design increasingly relies upon vegetation such as seagrass to reduce waves and currents, but modelling efforts usually assume static vegetation and do not consider the way in which the interactions between algae, light and nutrients can reshape seagrass canopy roughness throughout the year. For tropical shores, this is important because the light climate, epiphyte load, and dissolved nutrients vary with tides, monsoon month, and freshwater discharges. In this study, we bridge this gap with a robust systems dynamics module that is sensitive to real variability and captures stress resulting from competition among algae.

The model integrates three environmental drivers namely, photosynthetically active radiation (PAR), temperature, and dissolved nutrients consisting of nitrogen and phosphorous compounds. It predicts above- and below-ground biomass, shoot density, and nutrient allocation. Competition from algae is represented as epiphyte shading that filters light at the seagrass canopy. The framework is compact in design to preserve identifiability and allow for regular calibration.

Validation was conducted in two stages. First, model settings were applied to a temperate location, with method benchmarking and disciplined initialisation using configurations consistent with Carr (2012) and Kenov et al. (2013), and calibrated against Virginia Coast Reserve LTER data. Second, tropical applicability was assessed under Singapore conditions via continuous-flow mesocosms for a tropical species (i.e., Halophila ovalis) across varying light and nutrient conditions. Temporal trajectories of biomass and tissue nitrogen were compared with model predictions and supplemented by literature percent-cover and density datasets where available.

We separated the baseline seasonal pattern set by light and temperature from additional changes due to nutrient limitation, isolating regimes where nutrients, rather than physical condition, controlled growth. We then defined threshold bands and mapped these to habitat types to prioritise nutrient management or physical light restoration and to identify conditions where algal competition was most consequential.

The model serves as a quick stress test for seagrass-algae competition and can be coupled to hydrodynamic models for predicting spatial variations, with feedback from seagrass density and canopy traits to bed roughness and wave or current attenuation. The low-parameter design supports frequent calibration, uncertainty screening, and scenario testing design for nature-based tropical coastal defence against sea-level rise in response to climate change.

ABSTRACT. Floating photovoltaic (FPV) systems are increasingly deployed in coastal and near-shore environments where they are simultaneously influenced by surface waves and wind loads. These competing and interacting forcings generate coupled hydrodynamic and aerodynamic responses that remain insufficiently understood in existing FPV research, which typically treats the two effects separately. The purpose of this study is to establish a fully analytical framework capable of predicting the frequency-dependent dynamic behavior of a single FPV unit subjected to combined wave excitation and unsteady wind loading.

The key issue addressed is the unified treatment of wave-induced rigid-body motions and periodic aerodynamic forces, which are incorporated simultaneously into a single analytical dynamic formulation. Wave-induced surge, heave, and pitch motions modify the panel’s effective angle of attack, while aerodynamic lift and moment feed back into the system and alter its resonant characteristics. Capturing these mechanisms requires simultaneous treatment of hydrodynamic radiation–diffraction and unsteady thin-airfoil aerodynamics.

The methodology integrates two analytical components. Hydrodynamic forces are obtained using an eigenfunction-expansion solution of the Laplace equation, incorporating diffraction, radiation, added mass, and radiation damping. Aerodynamic forces are modeled using Theodorsen’s unsteady lift theory with a weak ground-effect correction, enabling representation of circulatory and non-circulatory contributions to lift and moment. These components are assembled into a frequency-domain equation of motion that yields the coupled surge–heave–pitch responses.

Results reveal that the FPV dynamics evolve continuously between two contrasting behaviors as wind speed varies. In the small-wind regime, the unsteady aerodynamic contribution is predominantly stabilizing, reducing resonance amplitudes and exerting little influence on wave scattering. As wind speed enters the higher range, the phase relation between aerodynamic lift and rigid-body motion changes, causing the aerodynamic force to act constructively with the hydrodynamic excitation. This interaction produces notable amplification in heave motion and alters the response frequencies, especially for shallow-draft configurations where hydrostatic restoring is weak.

The study concludes that analytical coupling of radiation–diffraction hydrodynamics and unsteady aerodynamic theory provides a powerful tool for understanding FPV dynamics. The framework offers physical insight and a foundation for design, optimization, and future extensions to large FPV arrays.

ABSTRACT. The linear-momentum actuator disc theory (LMADT) is one of the promising approaches for estimating tidal energy potential. In particular, when coupled with shallow water models, it has been widely used not only for tidal energy resource assessment but also for long-term prediction of hydrodynamic changes induced by the installation of tidal energy farms, thereby serving as a decision-support tool for the introduction, design, and planning of tidal power developments. In conventional LMADT-based approaches, tidal turbines are modeled as disk-type actuators, and a quartic equation is solved at the cell interfaces to extract the maximum available power, after which the resulting flow state (i.e., velocity and water depth) is imposed. This procedure requires relatively large computational cost in turbine-installed grid cells and increases numerical complexity. In this study, a decoupled framework is proposed in which the hydrodynamic response to tidal energy extraction and the estimation of tidal energy potential are separated through an energy-based variable transformation. The former is computed online during the simulation, while the latter is evaluated in a post-processing step. Specifically, instead of repeatedly solving the quartic equation at turbine-installed cell interfaces, an analytic formulation is imposed that directly maps the upstream flow in the extraction direction to the minimum-energy (critical) state, thereby significantly improving computational efficiency. The tidal energy potential is then extracted in post-processing from the pre- and post-transformation flow conditions using LMADT theory. In addition, the energy-based variable transformation employed in the numerical scheme is advantageous for preserving the well-balanced property of steady solutions, enabling more accurate and robust predictions of tidal energy potential.

ABSTRACT. Sediment entrainment around hydraulic structures is commonly predicted using time-averaged bed shear stress or Shields-type criteria. However, in the wake region of bluff bodies, strongly unsteady coherent turbulent structures dominate near-bed momentum exchange, potentially invalidating mean-flow-based approaches. This study investigates the role of wake-induced coherent turbulence in sediment entrainment downstream of a circular pile using high-resolution CFD–DEM simulations. The fluid phase was resolved using Large-Eddy Simulation (LES), while sediment particles were modeled explicitly using the Discrete Element Method (DEM) within a fully coupled CFD–DEM framework. This approach enables direct quantification of instantaneous fluid–particle interaction forces and particle trajectories at the grain scale. Spatiotemporal analyses of velocity fields and vortex structures reveal that sediment pickup events are closely correlated with coherent wake motions generated by vortex shedding. In particular, upward velocity bursts associated with coherent vortical structures promote particle lift-off, even under conditions where the mean bed shear stress remains below conventional entrainment thresholds. Particle trajectories exhibit complex behaviors, including intermittent upstream and lateral displacements, reflecting the organized and three-dimensional nature of the wake flow. The results indicate that sediment entrainment in wake-dominated regions is governed primarily by transient coherent structures rather than by time-averaged shear stress alone. Consequently, traditional formulations based solely on mean shear stress may underestimate local scour potential near hydraulic structures. Incorporating coherent turbulence effects into sediment transport modeling is therefore essential for improving predictions of structure-induced erosion and sediment dynamics in engineering applications.

ABSTRACT. Hydrological and hydraulic infrastructure resilience has traditionally been assessed through engineering-based indicators such as structural safety, hydraulic capacity, and recovery performance. While these indicators remain fundamental, recent climate-driven hydrological extremes reveal that infrastructure systems may perform within design thresholds while still producing severe environmental degradation, social vulnerability, and governance failure. This discrepancy highlights the need to reinterpret infrastructure resilience beyond purely physical performance. This paper proposes an ESG-based resilience assessment framework that integrates Environmental, Social, and Governance (ESG) dimensions into hydrological and hydraulic infrastructure evaluation, while preserving engineering rigor. The framework consists of three interconnected layers: (1) an engineering analysis layer based on hydrological modeling and hydraulic simulation under extreme conditions; (2) an ESG translation layer that systematically derives environmental, social, and governance indicators from engineering outputs; and (3) a decision-support layer linking ESG-informed resilience assessment to infrastructure planning, investment prioritization, and adaptive management. Rather than treating ESG as an external or qualitative assessment tool, the proposed framework conceptualizes hydrological and hydraulic analysis as a generator of sustainability-relevant value. The framework provides a conceptual bridge between physical system behavior and governance-oriented decision-making and offers a foundation for future quantitative indicator development and empirical validation. The study contributes to advancing resilience theory in hydraulic engineering and supports more comprehensive and sustainable water infrastructure management under climate uncertainty.

ABSTRACT. Vegetation in rivers acts as a resistance factor during floods, increasing the risk of inundation due to rising the water levels. At the same time, it causes micro-topographical changes such as the entrainment of fine-grained sediment in their wake-zones, providing valuable habitats for river ecosystems. Furthermore, the role of the vegetation as the resistance factor also includes flood control benefits, such as the energy dissipation of destructive flows, the velocity reduction near river structures, and attenuating discharge hydrograph with storing floodwater. The vegetation is therefore expected to serve as green infrastructure. In order to evaluate such diverse effects of the vegetation under various flood and river conditions, a reliable numerical analysis model is essential. However, the boundary of vegetation is too complex for fully resolved simulation with CFD models. Consequently, the influence of vegetation on fluid motion has been modelled through factors such as drag. Considering the subgrid-scale vegetation models and shallow river flows, a depth-integrated model is considered effective for simulating flows and sediment transports in vegetated channels. However, two-dimensional (2D) models cannot account for three-dimensional vortex (3DV) such as secondary flows. Therefore, even when incorporating turbulence models, the energy cascade with 3DV cannot be evaluated, presenting a challenge in adequately considering energy losses due to vegetation for 2D models. Furthermore, in destructive flows such as dam-break flows or tsunamis, the wavefront morphology changes significantly due to non-hydrostatic components and 3DV. In addition, existing drag models employ drag coefficient models for uniform flow, presenting challenges in their applicability to unsteady and non-equilibrium flow. While complex subgrid flows generated by vegetation have been modelled, sediment transport models incorporating vegetation have not yet been sufficiently investigated. In this study, to address the limitations of conventional numerical models with vegetation, we developed a novel depth integrated model for simulating interactions among three-dimensional flow-vegetation-sediment transport model. This model incorporates a drag model for non-equilibrium open channel flows and a sediment transport model based on turbulent tractive force, integrated into the enhanced depth integrated model capable of considering 3DV and non-hydrostatic pressure distribution. This study demonstrates the performance of the model to the application to simulate differences in wave profiles passing through vegetation groups at initial Froude numbers of the bores and bed topographical changes in a partially vegetated meandering channel.

ABSTRACT. Climate change and land-use change are increasing sediment yield from watersheds, accelerating in-channel deposition and thereby reducing river conveyance capacity, raising flood stages, and amplifying risks in river management. Accordingly, there is a growing need to mitigate sediment-related impacts through watershed–channel integrated sediment management based on Nature-based Solutions (NbS), which reduces sediment sources and alleviates depositional effects within channels. This, in turn, requires (i) the development of a physics-based, quantitative model capable of simulating sediment generation, transport, and deposition processes, and (ii) the establishment of an integrated assessment framework that can evaluate the applicability and effectiveness of NbS (land-cover change) scenarios. In this study, we developed a watershed-based model that jointly simulates rainfall–runoff–sediment yield and verified its applicability by evaluating the model performance for both discharge and sediment yield. We further designed NbS-based land-cover change scenarios for areas identified as having relatively high soil-erosion concern based on the Universal Soil Loss Equation (USLE), and quantitatively assessed scenario-dependent reductions in flood runoff and sediment yield. In addition, to represent soil organic carbon (SOC) dynamics and the attenuation of soil carbon conservation functions induced by soil erosion, we coupled a soil-carbon analysis module with the hydrological model and analyzed changes in CO₂ flux (soil respiration) and SOC before and after erosion events. The results indicate that the proposed model reproduced discharge and sediment yield in the testbed with overall strong performance. When applying the NbS-based land-cover change scenarios to high erosion-risk areas, flood runoff and sediment yield were reduced by up to 13.62% and 19.5%, respectively, demonstrating the effectiveness of NbS interventions. The simulated CO₂ flux captured short-term variability reasonably well except for certain periods. While rigorous validation of absolute SOC amounts was limited by the lack of long-term observations, the model indicated a subtle decreasing tendency in SOC associated with soil erosion. The integrated model developed in this study enables concurrent simulation of rainfall–runoff–sediment yield–carbon dynamics and can serve as a practical tool for quantitatively evaluating watershed hydrological, sediment, and carbon responses under various NbS and climate-change scenarios, thereby supporting the development of robust management and adaptation strategies.

Funding: This work was supported by Korea Environmental Industry&Technology Institute(KEITI) through R&D Program for Innovative Flood Protection Technologies against Climate Crisis Project, funded by Korea Ministry of Environment(MOE)(2022003460002)

ABSTRACT. Southeast Texas (SETx), United States, experiences coastal (storm surge), fluvial (riverine), and pluvial (rain-on-pavement) flooding, making it a hydrologically complex and hazard-prone region. The area also contains a disproportionate share of populations that are socially vulnerable to natural disasters, amplifying potential human and economic impacts. In this study, we assess coastal flood and hurricane impacts across SETx with a focus on hydroclimatic drivers of risk. Specifically, we use (i) hurricane downscaling techniques to better capture local-scale storm characteristics and (ii) flood frequency analysis (FFA) combined with stochastic storm transposition (SST) to evaluate changes in flood hazard under future climate conditions. This integrated approach allows us to extend the analysis from peak flow responses to flood extent, enabling a unique view and analysis of flood hazard and population flood exposure at the basin scale. Results indicate that hurricane-driven precipitation is likely to increase across SETx, contributing to increases in the magnitude and frequency of coastal flood hazards in the future. These findings highlight important implications for communities that are already socially vulnerable and provide stakeholders and emergency management agencies with a more comprehensive understanding of hurricane and flood impacts across the basin, supporting informed decision-making for flood risk assessment and management.

ABSTRACT. River ethics, which began to take shape in China in the early 2020s, proposes that rivers should be understood not merely as physical systems but as living entities, and potentially as beings with personhood, that possess intrinsic value. Within the broader domains of environment and water culture and of environment and water ethics, river ethics seeks to establish a morally grounded framework for understanding and reshaping human–river relationships. To examine the intrinsic similarity and inclusiveness of environment-water-river and culture-ethics, this presentation systematically shows the logical hierarchy of each key term. River ethics offers a hydrologically grounded, system-specific perspective, emphasizing the river as a dynamic eco-hydrological system rather than a passive conduit for water and sediment. This framework highlights the river as a self-organizing entity whose flow regime, geomorphic processes, sediment continuity, water quality, and biotic interactions collectively sustain riparian and watershed-scale ecological networks. By extending ethical consideration directly to the river itself, river ethics recognizes the intrinsic and connective values embedded in channel morphology, floodplain processes, groundwater exchange, and riparian vegetation dynamics. It advances a relational model centered on the harmonious coexistence of humans and rivers, acknowledging the limits of purely engineering-driven river control—particularly amid the intensifying uncertainties of climate variability in the twenty-first century—and advocating management approaches that respect both the physical dynamics of rivers and ecological integrity. River ethics has already been adopted in New Zealand and Peru, and, more recently, in the Colorado River, USA, as a new normative framework for human–river coexistence, with human representatives (spokespersons) appointed to act on behalf of the river.

ABSTRACT. Irrigation engineering heritage serves as a living carrier of civilization formed by human adaptation to nature and utilization of water resources. Its formation and evolution are the consequence of long‑term interaction and coupling between the natural environment and human society. Taking irrigation engineering heritage as the research object, this paper focuses on its coupling mechanism with natural and humanistic environments. It systematically analyzes the constraints and supporting effects of natural conditions such as hydrology, geomorphology and ecology on the layout and technical selection of irrigation projects, and discusses the influences of humanistic factors including historical governance, social institutions, farming culture and regional development on the functional evolution and value inheritance of the heritage. Based on literature interpretation and case analysis, this study reveals the internal mechanism of mutual adaptation and interaction between irrigation engineering heritage and natural‑humanistic environments. The results show that the natural environment provides the material foundation for irrigation engineering heritage, while the humanistic environment endows it with social functions and cultural connotations. Their coupling and coexistence constitute the core driving force for the existence and development of the heritage. Only by integrating natural conservation and activation of humanistic values can we realize the scientific protection, sustainable utilization and living inheritance of irrigation engineering heritage. This study can provide a theoretical reference for the protection of water heritage, the construction of water culture and the harmonious development between humans and water.

ABSTRACT. Under the Anthropocene, flood control policies increasingly confront the challenge of addressing not only hydrological hazards but also the complex entanglements of ecological processes, livelihood practices, and historically shaped landscapes. Conventional flood control frameworks, largely grounded in engineering and risk management, tend to treat flooding as an anomaly to be eliminated, often neglecting the ways in which riverine environments have long supported adaptive human–nonhuman coexistence. This study demonstrates how bio-cultural diversity can be integrated into flood control policy by adopting a multispecies perspective grounded in landscape history. The paper focuses on the lower Yoshino River basin in Tokushima, Japan, with particular attention to non-rice agricultural wetlands, especially lotus root fields, which function as artificial wetland environments. Since their introduction in the early twentieth century, lotus root fields have expanded through the entanglement of multiple historical events, including earthquakes, crop diseases, agricultural policy shifts, and changes in flood management. These landscapes are characterized by recurrent but shallow inundation that local farmers tend to tolerate. Such practices contribute to the maintenance of wetland-like conditions, supporting rich aquatic biota and attracting species such as the oriental stork, which has recently returned to the area. Methodologically, the study adopts a qualitative and interpretive approach combining landscape historical analysis, ethnographic fieldwork, and a multispecies framework. Historical documents, land-use records, field surveys, and interviews with farmers are used to reconstruct how flood management practices emerged from interactions among human livelihoods, nonhuman species, sediments, water flows, and infrastructures. Rather than framing flooding solely as a hazard, the analysis conceptualizes overflow as a generative process mediating ecological productivity, agricultural practices, and multispecies relations. The study argues that integrating bio-cultural diversity into flood control policy requires moving beyond the binary of protection versus risk. A multispecies perspective reveals that flood tolerance, controlled overflow, and seasonal inundation have historically played a crucial role in shaping resilient riverine landscapes. Flood control infrastructures and regulations can either disrupt or sustain these relations, depending on their engagement with local ecological dynamics and cultural practices. From this standpoint, the discussion resonates with rights of nature approaches, which reposition rivers as legal and ethical subjects within governance frameworks. In conclusion, the landscape history of the lower Yoshino River basin illustrates the potential of bio-culturally informed flood control to enhance resilience under the Anthropocene, offering a policy-relevant framework that aligns flood risk management with ecological diversity, livelihood continuity, and cultural heritage.

ABSTRACT. Water conservancy heritage embodies the profound historical wisdom accumulated through the long-term interactions between humans and rivers. Its value not only lies in remarkable engineering feats but also encompasses the dynamically evolving ethical conceptions of rivers. Based on an interdisciplinary framework integrating river ethics, heritage studies, and water history, this paper systematically analyzes the ethical narrative lineages inherent in global water heritage, with a special emphasis on China's abundant heritage. It further explores the potential transformation paths within the contemporary context of ecological civilization and sustainable development.

The analysis, through representative case studies, reveals a significant paradigm shift in river ethics. Chinese cases, such as the Dujiangyan Irrigation System and the Beijing - Hangzhou Grand Canal, demonstrate an evolving trajectory. In conjunction with international cases such as Subak Irrigation System in Indonesia and Sayamaike Reservoir in Japan, they collectively illustrate a global ethical transition from the earlier narratives of reverence, utilization, and conquest to an emerging narrative of coexistence, which recognizes the intrinsic value and rights of rivers.

Building upon this ethical re-framing, this study contends that the contemporary preservation and adaptation of water heritage should transcend the conventional approaches that solely focus on technical commemoration or tourism. Instead, it should be re-conceptualized as a cultural impetus and a practical platform for promoting sustainable river basin governance. The following key transformation paths are proposed: (1) Narrative reconstruction and value re-interpretation: Integrate ecological and social history to critically uncover indigenous ecological wisdom; (2) Spatial revitalization in synergy with ecological restoration: Incorporate heritage sites into broader watershed planning; (3) Innovate governance models and enhance community participation: Learn from the community co-governance mechanisms of traditional water conservancy societies, so as to empower multi-stakeholder co-governance and public participation in contemporary watershed management; (4) Digital technology empowerment for public engagement and global dialogue. Ultimately, this process facilitates the crucial transition of water heritage from “technological monuments” to active “sites for ethical reflection” and “platforms for ecological practice” contributing to the establishment of watershed life communities and the achievement of global sustainability goals.

ABSTRACT. It is necessary to improve the operation of existing dams in terms of both basin flood control and carbon neutrality. In order that, developing technologies for both “dam upgrading” to strengthen flood control functions and expand hydropower generation, and “restoration of sediment transport environments” to extend the life of dams and improve river and coastal environments are essential. In Japan, "Ecology and Civil Engineering” approaches to harmonize reservoir sedimentation management with sediment environment restoration has been discussed and several sediment replenishment, sediment sluicing and sediment bypassing projects which improve both reservoir life and river geomorphology below dams have been implemented. The damage caused by Typhoon Nabi in the Mimikawa River basin in 2005 resulted in significant sediment yield into mountainous areas, necessitating full-scale countermeasures to prevent sedimentation in reservoirs. For this reason, Miyazaki Prefecture formulated the Mimikawa River Basin Comprehensive Sediment Management Plan in 2011, which covers the entire basin, from the mountains to the sea. Sediment sluicing operation at the dams along the main stream of the Mimikawa River basin is a core component of this plan, and began in 2016 at Saigou and Ouchihara Dams, with Yamasubaru Dam being added in 2022, for a total of three dams. Upgrading work was carried out on the crest gates of Yamasubaru and Saigou Dams to ensure that water levels can be sufficiently lowered in the event of river flooding during typhoons. This presentation reports on the necessity of managing sediment at dams in the Mimikawa River basin, the validity of sediment sluicing as a sediment management measure, the formulation of sediment sluicing operation rules and confirmation of their effectiveness, advanced dam upgrading works to enable sediment sluicing, and changes in riverbed geomorphology below dams after sediment sluicing and positive environmental effects regarding habitat richness.

ABSTRACT. This study investigates the wave overtopping characteristics of horizontal composite breakwaters across various structural configurations through extensive two-dimensional hydraulic model experiments, ultimately proposing a new empirical formula for estimating overtopping discharge. Horizontal composite breakwaters frequently deviate from uniform designs. Some feature a caisson front fully armored with wave-dissipating blocks, others utilize configurations where portions of the armored section are replaced with rubble stones. This structural diversity often arises from practical site constraints, such as an insufficient supply of specialized armor blocks or the economic advantages of constructing a rubble mound beneath the caisson to obtain the cost-effectiveness. Consequently, these structures exhibit a wide range of wave-breaking behaviors influenced by the specific coverage ratio and armor layer thickness. While recent literature has proposed empirical formulas for calculating horizontal wave pressure on these composite structures, a critical research gap remains regarding wave overtopping discharge.

To bridge this gap and establish a framework that accounts for these configuration-dependent variables, systematic hydraulic experiments were performed in a two-dimensional wave flume. The experimental program focused on analyzing wave overtopping characteristics based on the armor configuration of the caisson front and the presence of a supporting rubble mound. Specifically, this research analyzed five distinct cross-sectional profiles: a baseline unarmored configuration, a fully and partially armored Tetrapod configurations, and rubble mound supported configurations with either full or partial armor layers of Tetrapod. Throughout the test program, wave overtopping discharge was measured under a wide range of wave heights and wave period conditions. By analyzing the data obtained, the study identifies the key governing parameters that influence discharge rates, providing the empirical basis necessary to predict overtopping for these types of structures.

ABSTRACT. Generally floating breakwaters and wave energy converters have been designed with a focus on either coastal protection or energy production. In this study, a new type of wave energy conversion concept based on a floating breakwater is proposed. The proposed structure combines a floating breakwater with a wave energy converter installed at its front. It consists of a semi-cylindrical buoy connected to a vertical plate, forming a rotating buoy at the front of the floating breakwater, where wave-induced motions are used to convert incident wave energy into electrical energy. The front-mounted buoy first generates electricity through its motion, and the floating breakwater then reduces the remaining wave energy. The floating breakwater uses slit sections with a porosity of 30% on the incident and transmission sides within the draft region, allowing wave reflection and energy dissipation. The rotating buoy with the attached vertical plate is designed to use not only the vertical but also the horizontal components of wave energy. Hydraulic model experiments were carried out to evaluate the wave energy reduction and the feasibility of power generation of the proposed structure. The experiments were conducted in a two-dimensional wave flume using a 1/30 scale model based on Froude similarity, under taut-leg mooring conditions with various wave periods and heights. The reflection and transmission coefficients of the floating breakwater and the generated electrical power were analyzed. In addition, the heave and pitch motion RAO (Response Amplitude Operator) of the front-mounted buoy were obtained through video analysis. Based on the obtained RAO, the power generation corresponding to the heave and pitch motions was separately evaluated. The experimental results show that continuous power generation was achieved for all experimental conditions, although the power output varied with wave period and height. The floating breakwater also reduced wave energy. The results of this study provide fundamental data for the development of a multifunctional coastal structure capable of both wave energy conversion and coastal protection.

ABSTRACT. Tsunamis can cause catastrophic damage to coastal communities when they strike the coast. This study aims to establish a rapid forecasting system for maximum tsunami height and arrival time using machine learning techniques. Imwon Port, located on the eastern coast of the Republic of Korea, was selected as the target site. To address uncertainties in seismic fault parameters and earthquake magnitude, a weighted logic tree approach was employed to generate a comprehensive tsunami dataset. Additionally, nine offshore observation points in the East Sea were considered to predict the maximum tsunami height and arrival time at Imwon Port. Machine learning models were constructed to predict maximum tsunami height and arrival time, and an inverse variance weighting method was introduced to combine the predictions from high-performing models. The estimates produced by the proposed machine learning framework showed good agreement with results from numerical simulations. The rapid predictions of tsunami characteristics enabled by this approach have the potential to reduce tsunami-related damages and save lives in vulnerable coastal areas.

ABSTRACT. This study presents an experimental investigation of wave reflection from a vertical oscillating water column (OWC) wave energy converter integrated with a conventional rubble-mound breakwater, with the aim of developing a practical empirical formula to estimate the reflection coefficient at the design stage. A series of two-dimensional hydraulic model tests was performed under a range of water depths and incident wave conditions representative of potential OWC installations. The results indicate that the OWC-equipped structure achieves lower reflection coefficients than a typical vertical breakwater, confirming its low-reflection behavior. The effects of non-dimensional internal free-surface oscillation amplitude, wave steepness, and relative water depth on the reflection coefficient were quantified. In addition, the internal water-column response was shown to be reasonably predicted using a one-dimensional theoretical model. Building on these observations, an empirical relationship incorporating the three variables was derived and demonstrated good agreement with the experimental data. The proposed formula is expected to support preliminary design and performance assessment of OWC-integrated coastal protection structures.

ABSTRACT. With rising sea levels and intensifying storms, accurate and rapid coastal wave forecasts are increasingly important for early warning and risk management. Physics-based spectral wave models such as SWAN provide reliable guidance but are computationally expensive, especially for real-time and ensemble forecasting in complex coastal domains. At the same time, recently proposed global deep-learning wave emulators typically require very large training datasets and operate at relatively coarse spatial and temporal resolutions, which limits their utility for local coastal applications that demand higher detail.

In this study, we develop a regionally specialized, boundary-informed deep-learning emulator that reconstructs hourly nearshore wave fields from SWAN hindcasts at high spatial resolution. The proposed architecture combines a UNet++ encoder–decoder with a ConvLSTM module to jointly capture spatial patterns and temporal evolution of coastal waves. The input includes atmospheric and oceanic forcings together with open-boundary directional wave spectra, bathymetry, and depth-averaged currents, allowing the model to account for wave transformation over complex topography and wave-energy inflow at the open boundary.

Sensitivity experiments indicate that explicitly incorporating directional spectra at the open boundary reduces mean absolute error by approximately 30–34% in our validation cases, highlighting the importance of boundary conditions in coastal wave emulation. The inclusion of depth-averaged currents and multi-scale supervision further improves predictive skill and spatial coherence of the reconstructed fields. Across independent validation periods, including four major typhoon events, the emulator reproduces SWAN-significant wave height and related fields with close agreement. Inference on modern GPUs achieves more than a 100-fold speed-up relative to SWAN, suggesting strong potential for near-real-time coastal wave forecasting and hazard mapping in operational settings.

<Acknowledgments> This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. RS-2024-00356663 and RS-2024-00444224).

ABSTRACT. In Korea, vertical and composite type structures are commonly applied using perforated wall structures. For the relatively deep water conditions, rubble-mound structures often become uneconomical because the required rubble mound width and rock volume increase substantially; consequently, the application of vertical concrete caissons has grown in recent years. Perforated-wall systems are applied to reduce wave reflection and wave loading and to enhance the hydraulic performance. Following the pioneering work of Jarlan (1961), lots of studies on perforated caissons have been carried out worldwide, with particular attention to the optimal chamber width. KICT (2000) experimentally investigated perforation ratio and reported that a 30% slit-wall perforation effectively decreases wave reflection. Kim and Lee (2013) further evaluated the slit length below the still water level relative to incident wave height, showing that a vertical-slit length of approximately 2.5 times the design wave height is effective for reflection reduction. This present study re-examines reflection performance by comparing vertical and transverse slit configurations. The experiments indicate that: (1) for the same perforation ratio and total perforated area, transverse slits with larger individual openings yield lower reflection coefficients; and (2) the optimal transverse-slit length below still water level is approximately 2–3 times the design wave height, consistent with trends reported for vertical slits.