Chorong Kim, Chung-Soo Kim, Chung-Soo Kim et al. · Tropical Cyclone Research and Review · 2021
Water level forecasting according to rainfall is important for water resource management and disaster prevention. Existing hydrological analysis is accompanied by difficulties in water level forecasting analysis such as topographic data and model parameter optimization of the area. Recently, with the improvement of AI (Artificial Intelligence) technology, a research using AI technology in the water resource field is being conducted. In this research, water level forecasting was performed using an AI-based technique that can capture the relationship between data. As the watershed for the study, the Seolmacheon catchment which has the rich historical hydrological data, was selected. SVM (Support Vector Machine) and a gradient boosting technique were used for AI machine learning. For AI deep learning, water level forecasting was performed using a Long Short-Term Memory (LSTM) network among Recurrent Neural Networks (RNNs) used for time series analysis. The correlation coefficient and NSE (Nash-Sutcliffe Efficiency), which are mainly used forhydrological analysis, were used as performance indicators. As a result of the analysis, all three techniques performed excellently in water level forecasting. Among them, the LSTM network showed higher performance as the correction period using historical data increased. When there is a concern about an emergency disaster such as torrential rainfall in Korea, water level forecasting requires quick judgment. It is thought that the above requirements will be satisfied when an AI-based technique that can forecast water level using historical hydrology data is applied.
Cited by 10Open access
Fi‐John Chang, Shenglian Guo · Water · 2020
The impacts of climate change on water resources management as well as the increasing severe natural disasters over the last decades have caught global attention. Reliable and accurate hydrological forecasts are essential for efficient water resources management and the mitigation of natural disasters. While the notorious nonlinear hydrological processes make accurate forecasts a very challenging task, it requires advanced techniques to build accurate forecast models and reliable management systems. One of the newest techniques for modelling complex systems is artificial intelligence (AI). AI can replicate the way humans learn and has the great capability to efficiently extract crucial information from large amounts of data to solve complex problems. The fourteen research papers published in this Special Issue contribute significantly to the uncertainty assessment of operational hydrologic forecasting under changing environmental conditions and the promotion of water resources management by using the latest advanced techniques, such as AI techniques. The fourteen contributions across four major research areas: (1) machine learning approaches to hydrologic forecasting; (2) uncertainty analysis and assessment on hydrological modelling under changing environments; (3) AI techniques for optimizing multi-objective reservoir operation; and (4) adaption strategies of extreme hydrological events for hazard mitigation. The papers published in this issue can not only advance water sciences but can also support policy makers toward more sustainable and effective water resources management.
Cited by 41Open access
Haeun Jung, Hyeongseop Kim, Jeongwon Lee et al. · 2026
ABSTRACTFloods are a major natural disaster that occurs suddenly, causing loss of life and significant socioeconomic damage, making proactive response through accurate forecasting essential. River water levels are a key indicator for determining flood occurrence, but forecasting accuracy varies depending on hydrological, meteorological, and watershed characteristics. In particular, small-scale rivers have higher water level variability compared to large-scale rivers, limiting their practical flood response capabilities. To overcome these limitations, this study aimed to develop an AI-based river water level forecasting model and improve its forecasting performance. The forecast model is configured to forecast river water levels three hours ahead using time-series data from the previous 24 hours as input, based on the forecast time point. Four locations among existing AI-forecast rivers where flood forecasting is difficult were selected, and input variables reflecting each location's hydrological, meteorological, and oceanographic characteristics were configured. As a result, this study confirmed a trend toward improved flood forecasting performance through the configuration of input variables tailored to each site's characteristics and the adoption of the latest AI models.AcknowledgmentsThis 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) and by the National Research Foundation of Korea(NRF) grant funded by the Korea government(MSIT)(RS-2025-00563294).
Cited by 0Open access
Huynh, Ngo Nghi Truyen · SPIRE - Sciences Po Institutional REpository · 2025
River discharge estimation and flood forecasting remain among the most compelling challenges in hydrological sciences, especially in ungauged basins and under climate change. Hydrological models are fundamental to flood warning systems, and enhancing their realism and accuracy is crucial. However, these models are inherently conceptual and subject to several limitations, including a lack of scale-relevant theory, intrinsic uncertainties in their structure and parameterization, and equifinality in parameter estimation, often due to sparse constraining data such as discharge measurements. In the age of artificial intelligence (AI) and big data, hydrological models can benefit from AI algorithms by exploiting massive amounts of information from heterogeneous datasets. This thesis introduces a set of innovative hybrid physics-AI approaches, seamlessly integrating process-based and AI-driven models, to improve hydrological modeling, with a particular focus on flash flood modeling in the French Mediterranean region. The suggested methods are implemented within SMASH (Spatially distributed Modeling and ASsimilation for Hydrology), a grid-based differentiable hydrological model developed for water research and operational applications. The key developments are: - Enhancing model accuracy through a multi-criteria calibration approach using multi-scale signatures; - Integrating neural networks into the forward model to estimate the transfer function between descriptors and parameters, enabling regionalization for ungauged sites; - Seeking optimal solutions using a Bayesian-weighted estimation method in the context of equifinality; - Reducing structural uncertainty by integrating neural networks to refine internal water fluxes within the model; - Embedding neural ordinary differential equations (neural ODEs) solved with an implicit numerical scheme suited for hybrid approaches. The results demonstrate that: - Multi-criteria calibration with hydrological signatures improves flood simulation performance while maintaining acceptable global scores; - Neural networks effectively learn complex non-linear functions between physical descriptors and conceptual parameters, exhibiting strong regionalization performance and better model controllability than linear regression in challenging calibration scenarios; - Stability of the optimal solution is enhanced by a Bayesian-weighted estimate of background values; - Neural networks, embedded for internal flux correction within a state-space system (from analytical or numerical ODE solutions), improve model accuracy and robustness. Detailed analysis reveals the nuanced behavior of hybrid models and their enhanced versatility, demonstrating the potential of hybrid physics-AI approaches to uncover laws and parameterizations in hydrology and beyond. These advancements have improved spatially distributed hydrological modeling by integrating AI into physical models, which can be applied to operational flood forecasting and support the development of new hydraulic approaches for runoff and flood modeling.
Cited by 0
Muneeza Azmat, Malvern Madondo, Arun Bawa et al. · 2023
Agriculture faces unprecedented challenges due to climate change, population growth, and water scarcity. These challenges highlight the need for efficient resource usage to optimize crop production. Conventional techniques for forecasting hydrological response features, such as soil moisture, rely on physics-based and empirical hydrological models, which necessitate significant time and domain expertise. Drawing inspiration from traditional hydrological modeling, a novel temporal graph convolution neural network has been constructed. This involves grouping units based on their time-varying hydrological properties, constructing graph topologies for each cluster based on similarity using dynamic time warping, and utilizing graph convolutions and a gated recurrent neural network to forecast soil moisture. The method has been trained, validated, and tested on field-scale time series data spanning 40 years in northeastern United States. Results show that using domain-inspired clustering with time series graph neural networks is more effective in forecasting soil moisture than existing models. This framework is being deployed as part of a pro bono social impact program that leverages hybrid cloud and AI technologies to enhance and scale non-profit and government organizations. The trained models are currently being deployed on a series of small-holding farms in central Texas.
Cited by 5Open access
G. Blandini, Simone Gabellani, Francesco Avanzi et al. · 2026
At the national level in Italy the FloodPROOFs hydrological forecasting chain is operational for flood forecasting, monitoring and emergency management. It is based on the physically based and spatially distributed hydrological model Continuum. While highly reliable, such modelling chains are often computationally demanding, posing limitations for rapid simulations and large-scale operational applications. To investigate whether artificial intelligence can support flood forecasting and support and integrate FloodPROOFs with comparable or better skill while significantly reducing computational costs, this study presents an AI-based framework for river water-level emulation designed for operational flood monitoring. The framework integrates a limited yet heterogeneous set of data sources typically available in real-time contexts, including topographic information derived from Digital Elevation Models, meteorological forcings from in situ measurements (precipitation and air temperature), and observed river water levels provided by the National System of Civil Protection and shared in myDEWETRA platform. Convolutional Neural Networks are employed to capture the nonlinear spatial and temporal interactions between terrain characteristics, atmospheric forcing, and hydrological response. The model is trained and fine-tuned using observed water-level time series, enabling the direct simulation of river stage dynamics and the detection of critical threshold exceedances relevant for civil protection warning procedures. The proposed framework operates at high spatial resolution over the Italian peninsula while maintaining low computational requirements, making it suitable for near-real-time applications at the centre of the work of the Italian Civil Protection. Its demonstrated generalization capability allows deployment across multiple spatial scales, from individual catchments to regional and national domains. Overall, the results highlight the potential of AI-driven emulators as complementary tools to traditional hydrological modelling chains, enhancing the efficiency and robustness of operational flood forecasting and decision-support systems for civil protection services.
Cited by 0Open access
Chitaranjan Dalai, Arun Elias, Goutham Kacheru et al. · 2025
Public safety in disaster operations relies heavily on accurate predictions which help minimize flood damages. The Long Short-Term Memory (LSTM) neural network demonstrates exceptional forecasting abilities in time-series applications thus making it an ideal choice for flood prediction operations. The presented research design implements a data-based flood prediction model which unites LSTM neural networks and collects hydrological information together alongside meteorological and topographical data to enhance forecasting precision. The predictive model uses IoT sensor along with satellite observation data to optimize hyperparameter adjustments which leads to better generalization outcomes. The forecasting performance is validated by three metrics including RMSE, MAE and R² which ensure dependable results. The model tackles three common limitations such as scarce data and excessive noise and calculation restrictions using transfer learning and regularization capabilities and cloud storage deployment. The experimental results show that LSTM maintains superiority over conventional statistical models and brings an effective scalable flood risk management system. The research investigates both hybrid AI systems and explainable AI approaches for enhancing AI readability and accuracy levels.
Cited by 1
Fi‐John Chang, Li‐Chiu Chang, Jui-Fa Chen · Water · 2023
The sustainable management of water cycles is crucial in the context of climate change and global warming. It involves managing global, regional, and local water cycles—as well as urban, agricultural, and industrial water cycles—to conserve water resources and their relationships with energy, food, microclimates, biodiversity, ecosystem functioning, and anthropogenic activities. Hydrological modeling is indispensable for achieving this goal, as it is essential for water resources management and mitigation of natural disasters. In recent decades, the application of artificial intelligence (AI) techniques in hydrology and water resources management has made notable advances. In the face of hydro-geo-meteorological uncertainty, AI approaches have proven to be powerful tools for accurately modeling complex, non-linear hydrological processes and effectively utilizing various digital and imaging data sources, such as ground gauges, remote sensing tools, and in situ Internet of Things (IoTs). The thirteen research papers published in this Special Issue make significant contributions to long- and short-term hydrological modeling and water resources management under changing environments using AI techniques coupled with various analytics tools. These contributions, which cover hydrological forecasting, microclimate control, and climate adaptation, can promote hydrology research and direct policy making toward sustainable and integrated water resources management.
Cited by 45Open access
Fenil Gandhi · Figshare · 2026
Accurate daily precipitation forecasting is essential for operational hydrology. This study evaluates six machine learning (ML) models for 1-day-ahead precipitation forecasting across 31 diverse stations in New Jersey (NJ), USA, using an integrated meteorological, remote sensing, and topographic dataset. Model performance was assessed for both general and 95th percentile extreme events. Ensemble models (random forest and XGBoost) consistently outperform other methods, particularly for high-impact extremes. Explainable AI (XAI) techniques, specifically SHAP, were used to interpret the models, revealing antecedent moisture conditions as dominant predictors. Crucially, the analysis demonstrates predictor importance is significantly modulated by land use/land cover (LULC), with urban areas showing heightened sensitivity to recent rainfall, reflecting the altered hydrological response of impervious catchments. Spatiotemporal analyses confirm predictive drivers are geographically structured. This research validates interpretable ML as a tool to improve forecast skill and enhance scientific understanding of heterogeneous hydrological processes.
Cited by 0Open access
Fenil Gandhi · Figshare · 2026
Accurate daily precipitation forecasting is essential for operational hydrology. This study evaluates six machine learning (ML) models for 1-day-ahead precipitation forecasting across 31 diverse stations in New Jersey (NJ), USA, using an integrated meteorological, remote sensing, and topographic dataset. Model performance was assessed for both general and 95th percentile extreme events. Ensemble models (random forest and XGBoost) consistently outperform other methods, particularly for high-impact extremes. Explainable AI (XAI) techniques, specifically SHAP, were used to interpret the models, revealing antecedent moisture conditions as dominant predictors. Crucially, the analysis demonstrates predictor importance is significantly modulated by land use/land cover (LULC), with urban areas showing heightened sensitivity to recent rainfall, reflecting the altered hydrological response of impervious catchments. Spatiotemporal analyses confirm predictive drivers are geographically structured. This research validates interpretable ML as a tool to improve forecast skill and enhance scientific understanding of heterogeneous hydrological processes.
Cited by 0Open access
Soo Young Kim, Seungho Lee, Kwang Seok Yoon · Earth · 2025
Global warming is recognized as a climate crisis that extends beyond a mere increase in the Earth’s temperature, triggering rapid and widespread climatic changes worldwide. In particular, the frequency and intensity of extreme rainfall events have increased in Korea and the Association of Southeast Asian Nations (ASEAN) region, leading to a significant increase in flood damage. The growing number of large-scale hydrological disasters underscores the urgent need for accurate and rapid flood-forecasting systems that can support disaster preparedness and mitigation. Compared with conventional physics-based forecasting systems, artificial intelligence (AI) models can provide faster predictions using limited observational data. In this study, a river water-level prediction model was constructed using real-time observation data and a long short-term memory (LSTM) algorithm, which is a recurrent neural network-based deep learning approach suitable for hydrological time-series forecasting. A repeated k-fold cross-validation technique was applied to enhance model generalization and prevent overfitting. In addition, water-level differencing was employed to convert nonstationary water-level data into stationary time-series inputs, thereby improving the prediction stability. Water-level observation stations in the Philippines, Indonesia, and the Republic of Korea were selected as study sites, and the model performance was evaluated at each location. The differenced LSTM model achieved a root mean square error of 0.13 m, coefficient of determination (R2) of 0.866, Nash–Sutcliffe efficiency (NSE) of 0.844, and Kling–Gupta efficiency of 0.893, thus outperforming the non-differenced baseline by approximately 17%. The repeated k-fold validation approach was particularly effective when the training data period was short or the number of input variables was limited. These results confirm that ensuring temporal stationarity and applying repeated cross-validation can significantly enhance the predictive accuracy of real-time flood forecasting. The proposed framework exhibits strong potential for implementation in regional early warning systems across data-limited flood-prone areas in the ASEAN region. Ongoing studies that apply and verify this approach in diverse hydrological contexts are expected to further improve and expand AI-based flood prediction models.
Cited by 0Open access
Qutada A.Ahmed, Mahmoud S Al-Khafaji, Dhiya Al-Jumeily · Water Conservation and Management · 2025
The precision of river flow management decisions depends on the applied techniques and decision support systems. In recent decades, Artificial Intelligence (AI) has emerged as a pivotal technology in decision-making. Determining the correct trends for using these applications based on a solid knowledge base requires a review of these applications and an assessment of their effectiveness in regions suffering from water scarcity and struggling with managing River’s flow, such as Iraq. This paper critically discussed the AI applications in the modeling and management of river flow with focusing on the Tigris River in Iraq as a case study. The review surveyed global and local literature on the use of AI in river systems, covering hydrology, hydraulics, river modeling, decision-making, water quality, water resources management, focus areas, and challenges, as well as, AI methods like ANNs, SVMs, and deep learning, which emphasize optimization, modeling, forecasting, and flood prediction. The findings indicated that the main challenges include data availability and quality (28%), model interpretability (22%) and system integration (20%). AI applications prioritize decision support systems (30%) and hydrological forecasting (25%), while irrigation efficiency (10%) is less emphasized. In river flow management, data availability (30%) and climate change (25%) are key issues, with flood forecasting (30%) and water quality monitoring (25%) being the most critical. The results showed that AI significantly enhances predictive capabilities, optimizes water distribution, and minimizes wastage, with critical applications in flood forecasting and water quality monitoring. The deployment of AI in Iraq needs a localized approach through improving data infrastructure enhancements. The Tigris River benefits from AI for sustainable river management even though the implementation poses various challenges to support climate change adaptation. The research shows AI has significant potential as a water resource solution for nations facing development challenges.
Cited by 0Open access
Jamal Hassan Ougahi, John S. Rowan · Scientific Reports · 2025
Understanding snow and ice melt dynamics is vital for flood risk assessment and effective water resource management in populated river basins sourced in inaccessible high-mountains. This study provides an AI-enabled hybrid approach integrating glacio-hydrological model outputs (GSM-SOCONT), with different machine learning and deep learning techniques framed as alternative 'computational scenarios, leveraging both physical processes and data-driven insights for enhanced predictive capabilities. The standalone deep learning model (CNN-LSTM), relying solely on meteorological data, outperformed its counterpart machine learning and glacio-hydrological model equivalents. Hybrid models (CNN-LSTM1 to CNN-LSTM15) were trained using meteorological data augmented with glacio-hydrological model outputs representing ice and snow-melt contributions to streamflow. The hybrid model (CNN-LSTM14), using only glacier-derived features, performed best with high NSE (0.86), KGE (0.80), and R (0.93) values during calibration, and the highest NSE (0.83), KGE (0.88), R (0.91), and lowest RMSE (892) and MAE (544) during validation. Finally, a multi-scale analysis using different feature permutations was explored using wavelet transformation theory, integrating these into the final hybrid model (CNN-LSTM19), which significantly enhances predictive accuracy, particularly for high-flow events, as evidenced by improved NSE (from 0.83 to 0.97) and reduced RMSE (from 892 to 442) during validation. The comparative analysis illustrates how AI-enhanced hydrological models improve the accuracy of runoff forecasting and provide more reliable and actionable insights for managing water resources and mitigating flood risks - despite the paucity of direct measurements.
Cited by 30Open access
David De León Pérez, Dariana Avila-Velasquez, Hector Macian‐Sorribes et al. · 2025
Seasonal hydrological forecasts are of critical importance for the effective management of water resources, particularly in complex and vulnerable basins such as the Mediterranean area which have both a deficitary water regimen and a high anthropogenic pressure. Nevertheless, inaccuracies in meteorological inputs can propagate through hydrological models, amplifying uncertainties in flow predictions. It is imperative to rectify these forecasts, whether meteorological or hydrological, to enhance prediction reliability and provide robust data for informed decision-making. This study proposes an advanced framework for seasonal hydrological forecasting that integrates raw and corrected weather forecasts with distributed hydrological modeling and sophisticated post-processing techniques to enhance flow prediction accuracy in the Júcar River in Spain as a representative case study of the Mediterranean area.The catchment hydrological model was implemented using the model TETIS v9.1 wich was calibrated and validated using observed records from 1981 to 2019 at seven control points. To do this, a split sample test was conducted using the period 2009–2019 for calibration, and the rest of the time series for validation (1981–2008). The process involved refining parameter maps to ensure good or acceptable performance at all control points. Meteorological data were sourced from the W5E5 dataset, downscaled to a 0.09° resolution using ERA5-Land. This improved the spatial and temporal resolution of the hydrological model. Once we have established an acceptable hydrological model,, the seasonal forecast hindcasts were evaluated using meteorological inputs from global forecasting systems, including ECMWF-SEAS5, CMCC_SPSv35, DWD_GCFS21, and MeteoFrance System8. To address uncertainties in the meteorological forecasts and their propagation to hydrological outputs, two complementary correction strategies were implemented. First, artificial intelligence(fuzzy logic) was applied to correct meteorological inputs before integration into the hydrological model, assuming errors originate solely from meteorological data and treating the hydrological model as a “perfect” simulator. Second, a hydrological error model was developed to identify and adjust discrepancies between simulated and observed flows, addressing systematic biases and errors in the hydrological simulation.The results demonstrated that forecasts based on corrected meteorological inputs exhibited significant accuracy improvements compared to those using unprocessed inputs. The hydrological error model further enhanced prediction reliability by mitigating systematic biases. These findings underscore the effectiveness of combining meteorological forecasts with AI-driven corrections to address uncertainties, thereby improving the robustness of seasonal hydrological predictions. This study highlights the potential for integrating advanced correction techniques into seasonal hydrological forecasting frameworks, offering a replicable methodology for other basins with similar complexities. The proposed framework enhances both the reliability and applicability of forecasts, ensuring their relevance for effective decision-making in complex hydrological systems as the Mediterranean area. The improved accuracy of these forecasts provides a sound scientific support for adaptive water resource management, particularly in the face of increasing climatic variability and environmental changes.Acknowledgments: This study was funded by the Colombian Ministry of Science, Technology, and Innovation (MINCIENCIAS) through the Call for Doctorates Abroad 885-2; by the Valencian Regional Government through the WATER4CAST 2.0 (CIPROM/2023/5) research project; and Spanish Ministry of Science and Innovation through the research project TETISPREDICT (PID2022-141631OB-I00). 
Cited by 0Open access
Aysu Ayar, Fatih Üneş, Bestami Taşar et al. · Aerul şi apa, componente ale mediului · 2025
With the decrease in water resources due to climate change, dam reservoir level estimation is important in terms of the construction, operation, design and safety of dams. In this study, average air temperature (T), relative humidity (SR), and precipitation (P) parameters were used for lake water level estimation. Thurmond Lake in McCormick County, South Carolina, USA was selected as the study area. 1286 daily data measured in real time between 2017-2023 were used as the study data. M5 Decision Tree (M5 Tree), Adaptive Neuro Fuzzy Inference System (ANFIS) and Artificial Neural Network (ANN) models were selected for lake water level estimation and model results were compared with real observation results. In the comparison of the prediction models, performance criteria such as coefficient of determination (R2), Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) were used. When the model results were examined; it was determined that artificial intelligence methods performed well in predicting the lake water level change.
Cited by 0Open access
Jane Alexander, Raphael Afonja, Sam Rubin · Abstracts with programs - Geological Society of America · 2016
Cited by 0
S.M. Sundaram, M.Lakshmi M.Lakshmi · PARIPEX-INDIAN JOURNAL OF RESEARCH · 2012
Cited by 6Open access
А.Ф. Петрушин, Yu.G. Yanko, А. В. Доброхотов et al. · RICE GROWING · 2024
Cited by 0Open access
Robby Yussac Tallar · 2024
The topic is quite interesting for predicting floods in the study area. The selected study area is Temerloh, Pahang, Malaysia. The authors need to put some explanations about the causes and also the historical floods in this area.
Cited by 0Open access
MOBOLAJI ADURAMO SODUNKE · 2024
Cited by 0Open access