Search Results (149)

The escalating frequency of extreme hydrological events under environmental uncertainty poses a severe socio-economic threat to floodplains such as the Brazilian Pantanal, the world’s largest tropical wetland. Mitigating dynamic flooding and drying cycles is highly challenging due to a critical scarcity of in situ monitoring, leaving flood risks poorly understood. To address these data gaps, this study presents an advanced deep learning forecasting framework that integrates multisource environmental data, fusing satellite-derived precipitation (CHIRPS) and global land data assimilation evapotranspiration (GLDAS) data with historical river gauge telemetry. Multi-layered neural network architectures were optimized and combined with progressive moving average filters (10− and 15−day windows) to capture the complex hydrometeorological patterns of the data-scarce Miranda River Watershed. The optimal deep learning configuration, utilizing a robust two-hidden-layer topology (15 and 60 neurons), consistently outperformed standard baselines. Although purely exogenous data blocks successfully minimized satellite noise and captured seasonal trends (NSE ≥ 0.92), structural underestimation of peak flows was observed. When incorporating the previous day’s streamflow (lag t−1) as a physical anchor, this limitation was noticeably alleviated, increasing both the Nash–Sutcliffe Efficiency (NSE) and Coefficient of Determination (R²) values above 0.99. While this performance surge is driven by the strong temporal persistence inherent to the autoregressive lag, it introduces an operational trade-off by restricting the forecast to a reactive 24 h window. In this regard, an evaluation of the operational forecast horizons revealed that the exogenous deep learning blocks maximize warning lead times, providing a vital tool for proactive civil defense and disaster risk reduction. Ultimately, this multisource framework establishes a methodological foundation for automated decision support systems, providing the high-accuracy streamflow forecasting capability required to support future flood mitigation frameworks. Full article

The floods triggered by the 29 October 2024 cut-off low in Valencia (Spain) expose a persistent challenge in disaster risk reduction: extensive meteorological and territorial data do not automatically become timely, trusted or actionable public guidance. This conceptual synthesis uses the Valencia event as a diagnostic case and reconstructs selected evidence on rainfall, hydrological escalation and alert timing to develop a data-driven framework spanning observation, modelling, impact assessment, communication, decision-making and post-event learning. Here, “data-driven” denotes an end-to-end governance and translation process, not the development of a new forecasting model. The framework integrates four dimensions: data governance, user-centred visualization, uncertainty communication and school-based education. It also introduces territorial translation as the link between impact forecasts and place-specific infrastructures, routines, vulnerabilities and responsibilities. Its novelty lies in connecting the Early Warnings for All pillars and impact-based, people-centred warning approaches with an explicit educational and territorial learning loop. Its practical contribution is a responsibility matrix, a minimum governance package and an implementation roadmap with indicators for latency, reach, comprehension and protective action. The framework is intended for adaptation, rather than statistical generalization, across Mediterranean and other fast-onset flood contexts. Improved forecasts remain necessary but insufficient: loss reduction requires interoperable records, accessible impact-based messages and inclusive educational programmes that convert scientific information into situated collective competence. Full article

Reservoir operation under hydrological uncertainty has become increasingly challenging under changing climate conditions. This study proposes a Probabilistic Dynamic Reservoir Operation Framework (PDROF) that integrates stochastic inflow modeling, Monte Carlo simulation, and dynamic rule extraction for adaptive reservoir management. Historical inflow records were transformed into stochastic inflow ensembles and propagated through reservoir operation simulations to generate reservoir storage trajectories under varying hydrological conditions. From these trajectories, a representative operational rule, referred to as the Most Likely Line (MLL), was extracted to characterize the dominant storage behavior of the system. The results demonstrate that conventional deterministic rule curves are constrained by predefined hydrological classifications and limited flexibility under variable inflow conditions. In contrast, the proposed framework effectively captures seasonal variability and propagates hydrological uncertainty throughout the operational cycle. Long-term simulation over a 23-year period resulted in a total spill volume of 17.74 million cubic meters (MCM), with spill events occurring in only 7 months, indicating improved operational robustness and storage stability. A real flood event in 2024 further demonstrated reductions of 47.42 MCM in spill volume and 3.46 MCM in reservoir storage compared with conventional operation. These improvements are attributed to the anticipatory storage behavior of the MLL-based operational rule, which preserves flood-buffer capacity prior to peak inflow periods and reduces the likelihood of uncontrolled spill events. The proposed framework provides a practical transition from deterministic reservoir operation toward uncertainty-aware and adaptive water resources management. The methodology is scalable to data-scarce and climate-sensitive regions and can be further extended through real-time forecasting and multi-objective optimization in future studies. Full article

Deterministic empirical formulas often fail to capture the epistemic uncertainties of dam failure mechanisms, leading to biased risk estimations. To address this, we propose the Probabilistic Hydro-Risk Forecasting System (PHRFS), a stochastic framework integrating Latin Hypercube Sampling (LHS) with HEC-RAS 2D hydrodynamic modeling. Applied to the Honghu-Erji Reservoir (Inner Mongolia), an ensemble of 60 stratified scenarios was generated based on a multi-model envelope of breach parameters and simulated over a $12.5\mathrm{m}$ ALOS PALSAR DEM using the Full Momentum Shallow Water Equations with the Eulerian–Lagrangian method (SWE-ELM). The SWE-ELM simulations reveal a heavy-tailed distribution of peak discharge (mean $2.65\times {10}^4{\mathrm{m}}^3/\mathrm{s}$; max $5.85\times {10}^4{\mathrm{m}}^3/\mathrm{s}$), significantly exceeding deterministic estimates. Sensitivity analysis identifies a physical dichotomy: breach depth ($D_b$) primarily controls flood magnitude ($r\approx 0.7$), whereas formation time ($t_f$) governs the arrival timeline. This temporal uncertainty propagates downstream, creating a distinct longitudinal gradient in SWE-ELM-derived warning time—ranging from rapid onset ($33.0\pm 19.3\min$) in proximal zones to a substantial lead time ($33.1\pm 4.2\mathrm{h}$) in the distal Hailar District. Consequently, a moderate coupling ($r=0.65$) emerges between economic loss and loss of life, while their spatial patterns remain strongly differentiated by warning time and exposure. Upstream settlements face high mortality risks due to insufficient evacuation windows, whereas the downstream urban center faces high economic exposure (∼1.57 billion CNY) but limited life loss (∼12.15 persons). These findings provide a scientific basis for differentiating emergency strategies, shifting from immediate life-saving in upstream reaches to asset protection in downstream areas. Full article

Flood-induced scour remains one of the leading causes of bridge failure, yet the chain linking flood hazard to bridge decisions is still commonly treated as a sequence of disconnected tasks. This review examines that chain using uncertainty as a unifying interpretive framework, synthesizing the recent literature on non-stationary flood hazard assessment, bridge-scale hydraulics, scour processes and predictive models, scour monitoring, monitoring-informed forecasting, structural vulnerability, and risk-informed decision-making. The review synthesizes the state of the art across all these stages of the chain, highlighting how the dominant uncertainty changes along it: climate and hydrologic variability upstream; model-form, sediment, and parameter uncertainty in scour prediction; measurement noise and inverse-inference uncertainty in monitoring; and threshold and consequence uncertainty in closure, retrofit, and network-level decisions. Although major advances have been achieved in probabilistic modelling, machine learning, hybrid physics-informed methods, and multimodal sensing, most published frameworks still transfer deterministic outputs from one stage to the next. As a result, uncertainty is rarely propagated consistently to the decision level. The main value of this review lies in making the chain’s weak interfaces explicit, in showing how uncertainty propagation can serve as a unifying framework across otherwise disconnected literatures, and in identifying which methodological directions are most promising for connecting prediction, monitoring, and decision support into a coherent end-to-end probabilistic chain supporting climate-resilient bridge management. Full article

Reliable estimates of future precipitation are essential for adaptive water management in large river basins. This study presents a machine-learning approach that combines six CMIP6 models to examine precipitation changes in the Yangtze River Basin. ERA5 monthly precipitation for 1979–2025 served as the reanalysis reference. The random forest model incorporated individual model outputs, ensemble statistics, geographic variables, and monthly cyclic terms. It was trained with data from 1979–2009, evaluated for 2010–2014, and then applied to the period 2015–2099 under SSP1-2.6, SSP2-4.5, and SSP5-8.5. Compared with the simple multi-model mean, the proposed method showed better agreement with ERA5 and generally smaller reconstruction errors during the validation period. Annual precipitation is projected to increase under all three pathways, with the largest increase under SSP5-8.5. Precipitation remains concentrated from May to August, while spring totals and intra-annual variability increase more clearly under high-emission conditions. Mean precipitation remains highest in the humid middle and lower reaches, while the magnitude and significance of future trends vary across the basin. Inter-model spread remains greater than the differences among emission pathways and reaches 85.92 mm under SSP5-8.5 during 2071–2099. These results represent uncertainty-aware climate estimates rather than verified forecasts. They can support flood-risk assessment, reservoir planning, and adaptive water management in the Yangtze River Basin. Full article

Reliable short-term hydropower forecasting is essential for dispatch planning and early electricity procurement in snowmelt-influenced power systems. This study develops a leak-free operational forecasting framework using quality-controlled hourly generation and hydro-meteorological records from eight hydropower plants in Kazakhstan. Two tasks are addressed: deterministic multi-step forecasting for D+1–D+7 and uncertainty-aware envelope forecasting for D+8–D+14 using MIN and Q90 targets. The benchmark uses Persistence as the primary baseline, against which RIDGE, SARIMAX, Random Forest, HistGradientBoosting, MLP, and LSTM are compared using Nash–Sutcliffe efficiency (NSE), root mean squared error (RMSE), and mean absolute error (MAE). For D+1–D+7, the results reveal strong cross-station heterogeneity and the expected decline in skill with increasing lead time. In the aggregated comparison, SARIMAX achieves the highest mean NSE at D+1 (0.903), while RIDGE becomes strongest by D+7 (0.625), both outperforming Persistence (0.534 at D+7). At the station level, SARIMAX performs best for Kapch, Kask, Moin, Bukh, and Ustk, RIDGE is best for Shar and Lenin, and LSTM is best for Shulb. The strongest stations, Kapch and Kask, reach mean NSE values of 0.941 and 0.933, respectively, whereas Ustk and Bukh remain the most difficult cases. A central methodological contribution is a flood-sensitive switched hybrid strategy for Ust-Kamenogorsk based on an observed-generation high-flow window selected by a regime-score procedure. This strategy improves robustness at medium lead times: for SARIMAX, NSE increases from 0.587 to 0.739 at D+2 and from 0.161 to 0.559 at D+7, while for RIDGE, NSE increases from 0.549 to 0.701 at D+2 and from 0.109 to 0.435 at D+7, together with substantial RMSE and MAE reductions. For D+8–D+14, envelope forecasting remains informative, but model ranking becomes target-dependent: SARIMAX and RIDGE provide the strongest mean performance for MIN (0.664 and 0.658), whereas LSTM and RIDGE are strongest for Q90 (0.746 and 0.743). Overall, the results show that hydropower forecasting in Kazakhstan is best approached as a station-wise, regime-aware, and horizon-specific problem. Full article

Accurate multi-horizon forecasting of reservoir water levels is essential for effective water resource management and flood risk mitigation. While deep learning models have demonstrated strong predictive capabilities, they often struggle to adequately represent uncertainty and extreme hydrological events, particularly at longer forecast horizons. This study proposes four variants of a Conv1D–LSTM–Temporal Attention (CLTA) architecture for probabilistic multi-horizon forecasting, differing exclusively in loss function design. The models incorporate non-crossing constraints, horizon-aware weighting, and event-aware weighting to address key limitations of standard quantile regression approaches. All models are trained on hourly water level data from May 2021 to October 2022 and evaluated on a fully unseen dataset spanning December 2022 to May 2023. The results demonstrate that horizon-aware weighting achieves the lowest average RMSE (0.0149) and the most stable performance across forecast horizons on unseen data, while event-aware weighting improves representation of extreme hydrological events and achieves the highest coefficient of determination ($R^2=0.9961$). However, a controlled experiment further reveals that model performance is strongly influenced by the data partitioning strategy, even when architecture and loss formulation are held constant. Overall, the findings indicate that loss function design, in interaction with data partitioning strategy, is a key contributing factor to model performance in deep learning-based hydrological forecasting. A Multi-Criteria Decision Analysis (MCDA) framework identifies the horizon-weighted model as the most robust general-purpose solution, while the event-aware model is preferable for applications focused on extreme event representation. These results highlight the importance of integrating domain knowledge into both model design and evaluation strategy, offering a scalable and computationally efficient alternative to increasing architectural complexity. Full article

Accurate flood simulation in regulated, low-lying river basins is crucial for forecasting and risk mitigation, but performance depends strongly on whether models represent floodplain hydrodynamics and human regulation. This study evaluates three coupled hydrological–hydrodynamic schemes in the Huai River Basin upstream of Bengbu Station using identical meteorological forcing and VIC-generated runoff: (I) a linear routing scheme (VIC–Routing), (II) a natural hydrodynamic scheme (VIC–CaMa-Flood), and (III) an extended hydrodynamic scheme that incorporates reservoir regulation and levee effects (VIC–CaMa-Flood with Dam). Results reveal clear spatial differences in scheme suitability. The linear routing scheme performs best in upstream reaches, with NSE and KGE generally exceeding 0.81, but tends to overestimate peak discharge in downstream lowland sections. Incorporating hydrodynamic processes and regulation representation further reduces peak flow bias. Scheme III achieves the most consistent downstream improvement, particularly for high flows (>2000 m³/s), with NSE exceeding 0.80 in long-term simulations and improved agreement with satellite-driven inundation patterns. However, simplified reservoir operating rules can increase uncertainty in water level dynamics. During the 2020 plum rain flood, Scheme II yielded more accurate water levels in some reaches, suggesting that generalized operation rules may introduce compensating errors even when discharge accuracy improves. Overall, reliable flood simulation in well-managed basins requires an explicit representation of both floodplain hydrodynamics and regulation, and scheme selection should be guided by the dominant controls along the river network. Full article

Vibration-assisted water flooding (VA-WF) can improve sweep efficiency. However, unclear macro-scale mechanisms limit its wider adoption in heavy oil reservoirs. This study combines previous sandpack experiments with two-dimensional Volume-of-Fluid (VOF) simulations to show how vibrations reshape permeability fields and, in turn, pressure and production behaviour. Heavy oil sandpacks were water-flooded under conditions of no vibration and 2 Hz and 5 Hz axial excitation. Measured injection pressure histories and oil production were used to calibrate a VOF model in which absolute permeability follows a log-normal distribution with directional anisotropy. Only when axial and radial permeabilities were assigned a negative local correlation did the model reproduce key observations: secondary pressure spikes, irregular viscous-fingering morphologies, delayed production drops, and variability in cumulative recovery. Parameter sweeps quantify the sensitivity of VA-WF performance to the variance and correlation of the permeability field, and multiple runs estimate the variability in outcomes introduced by stochastic heterogeneity. This study proposes a transferable workflow—comprising sample testing, parameter inference, and probabilistic simulation—to screen excitation conditions and forecast VA-WF performance prior to field implementation, enabling operators to optimize vibration frequency based on reservoir-specific permeability characteristics and to anticipate production variability under uncertainty. These results highlight the dominant factors affecting swept volume and oil recovery, supporting data-driven decision making in VA-WF projects. Full article

This paper presents a comprehensive real-time forecasting and mapping cycle of a regional flood event, encompassing quantitative precipitation forecasting, runoff production and routing, and inundation mapping. The objective of this study is to highlight the significant uncertainties inherent in each step of the fully automated cycle, despite the utilization of state-of-the-art models and remote sensing technologies. The case study focuses on a significant flood event that occurred in the Turkey River and Upper Iowa River, in rural Iowa, United States, resulting in localized damage and disruption to several small communities. The novelty of this study is that it demonstrates the limited utility of satellite-based remote sensing in the absence of other forecasting and mapping system elements, emphasizing the need for the timely integration of information from diverse sources to accurately forecast and map floods. To achieve this, we assembled and analyzed precipitation data from weather radars, streamflow estimates derived from river stages and rating curves, and cross-sectional data from river channels to characterize the movement of the flood wave. These data were integrated into hydrologic and hydraulic models to generate flood inundation estimates for the more severely affected areas. Remote sensing imagery was obtained and used as reference to assess the accuracy of the modeled inundated areas. Our findings illustrate that, despite the increasing availability of satellite data sources, there are still significant limitations to tracking inundation using satellite remote sensing, particularly for medium-sized basins. Flood modeling processes are not merely complementary to satellite-based flood estimation, but essential for comprehensive flood risk assessment. Full article

Floods pose significant risks to human lives, infrastructure, and the environment. Timely and accurate flood forecasting plays a pivotal role in mitigating these risks. This study proposes a Digital Twin proof-of-concept framework aimed at improving flood forecasting and validated its effectiveness through a pilot study of the 2021 flood event in Luxembourg. The baseline forecasting method combines GloFAS ensemble streamflow forecasts with a high-resolution flood hazard datacube generated using a LISFLOOD-FP hydrodynamic model and then averaging among the member forecasts. To dynamically update the flood forecasts and improve their accuracy, the framework integrates satellite-based Earth observations (EOs)—specifically Sentinel-1-derived flood probability maps from the Global Flood Monitoring service—via a particle filter-based data assimilation (DA) process. As such, the simulations with more coherence with the observed Sentinel-1-derived flood probability maps are prioritized. This results in a Digital Twin capable of delivering daily flood depth forecasts, at detailed spatial resolution, up to 30 days ahead, with reduced prediction uncertainty. Using the 2021 flood event, we evaluate the performance of the Digital Twin in assimilating EO data to refine hydraulic model simulations and issue accurate flood forecasts. Although certain challenges persist—particularly the difficulty in quantifying the error structure of GloFAS discharge forecasts—the proposed approach demonstrates clear improvements in forecast accuracy compared to open-loop simulations. As a result, the approach reduces water level prediction errors by an average of 15–33% and increases the Nash–Sutcliffe Efficiency of discharge predictions by approximately 15–36%. Future work will aim to refine the flood hazard datacube and advance the characterization and modeling of uncertainties associated with both GloFAS streamflow forecasts and Sentinel-1-derived flood maps, thereby further enhancing the system’s predictive capability. Full article

Probabilistic streamflow models play a pivotal role in quantifying hydrological uncertainty and form the backbone of modern risk management strategies for flood and drought forecasting, water allocation planning, and the design of resilient infrastructure. Unlike deterministic approaches that yield single-point estimates, these models provide a spectrum of possible outcomes, enabling a more realistic assessment of extreme events and supporting informed, sustainable water resource decisions. By explicitly accounting for natural variability and uncertainty, probabilistic models promote transparent, robust, and equitable risk evaluations, helping decision-makers balance economic costs, societal benefits, and environmental protection for long-term sustainability. In this study, we introduce the bounded half-logistic distribution (BHLD), a novel heavy-tailed probability model constructed using the T–Y method for distribution generation, where T denotes a transformer distribution and Y represents a baseline generator. Although the BHLD is conceptually related to the Pareto and log-logistic families, it offers several distinctive advantages for streamflow modeling, including a flexible hazard rate that can be unimodal or monotonically decreasing, a finite lower bound, and closed-form expressions for key risk measures such as Value at Risk (VaR) and Tail Value at Risk (TVaR). The proposed distribution is defined on a lower-bounded domain, allowing it to realistically capture physical constraints inherent in flood processes, while a log-logistic-based tail structure provides the flexibility needed to model extreme hydrological events. Moreover, the BHLD is analytically characterized through a governing differential equation and further examined via its characteristic function and the maximum entropy principle, ensuring stable and efficient parameter estimation. It integrates a half-logistic generator with a log-logistic baseline, yielding a power-law tail decay governed by the parameter $\beta$, which is particularly effective for representing extreme flows. Fundamental properties, including the hazard rate function, moments, and entropy measures, are derived in closed form, and model parameters are estimated using the maximum likelihood method. Applied to four real streamflow data sets, the BHLD demonstrates superior performance over nine competing distributions in goodness-of-fit analyses, with notable improvements in tail representation. The model facilitates accurate computation of hydrological risk metrics such as VaR, TVaR, and tail variance, uncovering pronounced temporal variations in flood risk and establishing the BHLD as a powerful and reliable tool for streamflow modeling under changing environmental conditions. Full article

Flood events driven by climate variability and change pose significant risks for socio-economic activities in the Awash Basin, necessitating advanced forecasting tools. This study benchmarks five deep learning (DL) architectures, Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Bidirectional LSTM (BiLSTM), and a Hybrid CNN–LSTM, for daily discharge forecasting for the Hombole catchment in the Upper Awash Basin (UAB) using 40 years of hydrometeorological observations (1981–2020). Rainfall, lagged discharge, and seasonal indicators were used as predictors. Model performance was evaluated against two baseline approaches, a conceptual HBV rainfall–runoff model as well as a climatology, using standard and hydrological metrics. Of the two baselines (climatology and HBV), the climatology showed limited skill with large bias and negative NSE, whereas the HBV model achieved moderate skill (NSE = 0.64 and KGE = 0.82). In contrast, all DL models substantially improved predictive performance, achieving test NSE values above 0.83 and low overall bias. Among them, the Hybrid CNN–LSTM provided the most balanced performance, combining local temporal feature extraction with long-term memory and yielding stable efficiency (NSE ≈ 0.84, KGE ≈ 0.90, and PBIAS ≈ −2%) across flow regimes. The LSTM and GRU models performed comparably, offering strong temporal learning and robust daily predictions, while BiLSTM improved flood timing through bidirectional sequence modeling. The CNN captured short-term variability effectively but showed weaker representation of extreme peaks. Analysis of peak-flow metrics revealed systematic underestimation of extreme discharge magnitudes across all models. However, a post-processing flow-regime classification based on discharge quantiles demonstrated high extreme-event detection skill, with deep learning models exceeding 89% accuracy in identifying extreme-flow occurrences on the test set. These findings indicate that, while magnitude errors remain for rare floods, DL models reliably discriminate flood regimes relevant for early warning. Overall, the results show that deep learning models provide clear improvements over climatology and conceptual baselines for daily streamflow forecasting in the UAB, while highlighting remaining challenges in peak-flow magnitude prediction. The study indicates promising results for the integration of deep learning methods into flood early-warning workflows; however, these results could be further improved by adopting a probabilistic forecasting framework that accounts for model uncertainty. Full article

Background: Despite advances in monitoring and forecasting systems, natural disasters continue to cause significant human losses. During the response phase, fast decisions are required to allocate limited resources, particularly rescue helicopters, which play a key role in reaching inaccessible areas. However, helicopter allocation involves trade-offs between efficiency and operational safety under uncertain conditions. Methods: This study proposes a decision-support methodology based on Multi-Criteria Decision Analysis (MCDA) for allocating rescue helicopters during disaster response. The approach integrates Value-Focused Thinking (VFT) and Multi-Attribute Value Theory (MAVT) to structure objectives, assign weights, and evaluate alternatives using criteria related to mission safety, response time, and expected number of rescued victims. The method is illustrated through a simulated flood response scenario in a Brazilian regional context. Results: The results show that the model allows decision-makers to compare allocation scenarios and to make explicit the trade-offs between operational efficiency and safety. The application indicates that small reductions in efficiency may lead to relevant gains in operational safety, particularly under adverse weather conditions. Conclusions: The proposed approach provides a transparent and traceable structure for supporting helicopter allocation decisions during disaster response. It contributes to more consistent decision-making in critical operations, especially in contexts characterized by uncertainty and time pressure. Full article