Accurate rainfall forecasting is vital for reservoir operation, shifting focus from minimizing missed detections in flood season to reducing false alarms post-flood. Conventional Multi-Model Ensemble (MME) methods use static weights, ignoring forecast skill variations across rainfall intensities. This study proposes MME-MID, a novel framework using Normalized Average Mutual Information Decomposition to dynamically weight forecasts by categorized rainfall intensities. Two decomposition approaches are applied: (1) Uncertainty decomposition prioritizes higher-accuracy categories to reduce post-flood false alarms; (2) Information decomposition enhances detection of high-intensity events during flood season. Validation in Dahuofang and Huanren basins shows MMEMID reduces miss ratio by 5.5 % in flood season and false alarm ratio by 11 % post-flood versus conventional MME, without increasing overall error, significantly improving decision reliability.

Hydrological regimes in complex river basins are governed by heterogeneous hydrological interactions, including driver-to-response interactions induced by multi-source inflows and response-to-response interactions within basin-scale hydrological processes. These interactions are essential for guiding scientific river-basin regulation, while their systematic quantification remains challenging, constrained by computation-intensive modeling and scenario-limited analysis. To this end, this study develops an Interpretable Spatio-Temporal Hydrological Interaction Learning (InSTIL) framework, which combines LSTM-based temporal encoding, graph-based spatial propagating and Shapley Additive Explanations, to explicitly capture and quantify the hydrological interactions. Applied to the Three Gorges Reservoir (TGR) and Dongting Lake (DTL) system, InSTIL delivers robust simulations of lake water levels with an average RMSE of 0.425 m, consistently outperforming LSTM and GCN benchmarks, particularly under extreme conditions. Furthermore, interaction interpretation reveals that: (1) For driver-toresponse interactions between rivers and the lake, their spatial patterns from different rivers are governed by inflow locations and runoff volumes, with TGR releases exerting the dominant influence (contributing 32.89%48.58% across individual sites) and tributary Yuan River ranking second. Meanwhile, these interactions appear unimodal over time, showing the strongest daily impacts at 1-3 days before lake's responses. Their multi-day cumulative effects inform an effective regulation window of 7-15 days. (2) For response-to-response interactions within the lake, the heterogeneous response timing and magnitude across DTL reveal the weak hydraulic connectivity in the south sub-lake and the presence of a cross-sublake hydraulic pathway. Overall, InSTIL provides a transferable paradigm for robust hydrological simulations and systematic interaction interpretation, offering practical insights for sustainable river-basin management.

The spatial resolution of hydrological modeling is a critical factor affecting flood simulation accuracy, especially in large watersheds characterized by complex watershed characteristics. However, its influence on the accuracy of hourly flood simulations at both watershed outlets and internal locations remains insufficiently understood, hindering rational spatial-resolution selection for large-scale flood forecasting. This study evaluates hourly flood simulations across five spatial resolutions (1, 3, 5, 10 km, and sub-watershed) at the watershed outlet and multiple internal stations in the Jialing River Basin, China (157 041 km(2)). An XGBoost-based model is employed to identify flood characteristics sensitive to spatial resolution and to quantify their nonlinear effects on simulation accuracy. Based on these relationships, spatial-resolution recommendations are derived for different flood-characteristic categories, and the effectiveness of spatial refinement under coarse rainfall inputs is examined. Results show that spatial refinement markedly improves simulation accuracy at internal locations but yields only marginal gains at the watershed outlet. Watershed area is identified as the dominant factor governing resolution sensitivity, while rainfall characteristics and underlying-surface properties exert strong nonlinear influences. Fine grids (1-3 km) are most effective under flood conditions with strong nonlinearity, but their advantages diminish rapidly as rainfall inputs become coarser, indicating that increased spatial resolution cannot compensate for insufficient rainfall information. Overall, these findings advance current understanding of spatial-resolution effects on hourly flood simulations and provide practical guidance for spatial-resolution selection in large-watershed modeling.

Quantifying the complex spatial-temporal correlations and generating representative high-dimensional coupled scenario sets are essential for the robust planning and risk assessment of large-scale hybrid energy systems (HESs). Although numerous models have been developed for this purpose, as the number of plants scales up to hundreds, existing approaches suffer from the curse of dimensionality, often resulting in high computational burden, posterior collapse, and distributional oversmoothing. To address this gap, this paper proposes a Spatial-Clustering Conditional Variational Auto-Encoder (SC-CVAE) framework, which employs spatial clustering to decompose the high-dimensional global problem into tractable subproblems and integrates adaptive deep networks to accurately capture high-dimensional spatiotemporal complementarity. Case studies on the Yalong River energy base, featuring massive wind and solar integration, demonstrate that SC-CVAE reduces global spatial correlation error by 56% compared to the Independent Baseline, while achieving a 2.4-fold computational speedup over the monolithic Global Baseline. Crucially, by mitigating posterior collapse to alleviate oversmoothing effects inherent in high-dimensional VAEs, the proposed framework improves the capture rate of high-impact extreme events by 3.4-fold and reduces the Energy Score error by 65%. This high-fidelity reconstruction of tail characteristics provides a more reliable basis for identifying supply-deficit risks in basin-wide HESs. The proposed framework enables scalable and high-fidelity generative modeling, establishing a robust methodology for stochastic optimization and long-term security assessments in the global transition toward decarbonized power systems.

In actual reservoir operations, water levels are often below the flood-limited water level, providing additional flood control storage and enabling the reservoirs to store more floodwater. This allows other reservoirs to reduce flood control storage through storage substitution, raising their water levels to enhance conservation benefits. The key question is how much the water level can raise, i.e., what is the substitution relationship of flood control storages. To address this issue, this study develops a flood control storage substitution model to utilize additional flood control storage without increasing the system's flood control risk. The cascade hydropower system of the Wudongde, Baihetan, Xiluodu, Xiangjiaba, and Three Gorges Project on the Yangtze River, China, is taken as a case study. Results indicate that the water level of the Three Gorges Project can be raised to 153 m when the operating water level of XLD and BHT are at dead water levels. The substituted ratio is less than 1.0 due to the discharge of the Three Gorges Project is reduced by substitution. The substituted ratio and upper bound of the substituted flood control storage are related with the volume of additional flood control storage, flood magnitudes and hydrographs, and the flood composition. In real-time operations, the storage substitution relationship can be used to adjust the operating water levels and increase power generation without increasing flood risk.

Ecological operation of reservoirs to control the nutrient transport has been a new hotspot problem in the recent reservoir operation research. Previous studies have focused more on optimization and mechanisms, neglecting the practicality of real-time operation rules, which makes it difficult for reservoir managers to conduct real-time ecological operation of reservoirs for nutrient transport, and such real-time operation is more difficult due to the hydraulic connections between upstream and downstream reservoirs in cascade reservoirs. Therefore, this paper, taking Lancang River cascade reservoirs as an example, combining total phosphorus (TP) mass balance with the multiobjective operation model, which is based on operation charts and optimized by NSGA-II, proposes a new operation chart, i.e., the TP export operation chart (TOC), by adding three types of parameters to the conventional operation chart (COC), to guide real-time TP export ecological operation of reservoirs, thereby reducing TP retention in reservoirs and making the real-time ecological operation for cascade reservoirs easier. The multiobjective optimization results allow decision-makers to better balance power generation and TP export by selecting appropriate solutions on the Pareto front. First, TOC-specific parameters are derived from the comparison between Pareto front results taking COC and operation processes as decision variables. Then, TOC is taken as decision variables to optimize the reservoir operation, and the rationality of the optimization results is analyzed. Finally, typical solutions are selected from the TOC Pareto solution set to simulate the long-term operation and examine the practical application effect of TOC. The results indicate that TOC can greatly compensate for the shortcomings of the COC and significantly improve the TP export amount. When applying the optimal solutions for TP export, TOC can reduce the TP retention by 5% on average compared to COC, because TOC can guide the unique characteristics of operation processes that are beneficial for TP export through the extra discharge operations defined by new parameters, which COCs cannot achieve. The simulated operation results and optimized operation results of TOC are in good agreement, which means the TOC optimization results have good applicability. This paper made up for the lack of research of real-time operation rules of TP export operation of reservoirs and has important guiding significance for the formulation of rules for reservoir management.

Frequent and severe floods in the transboundary river pose a great threat to the lives and well-being of people in riparian countries, which urgently requires basin-wide s joint flood control operation of those reservoirs. However, coordinating a large number of reservoirs presents challenges in both solution development via complex models and practical implementation. To resolve those challenges, we identify key reservoirs for joint operation, which play essential roles in basin-wide flood control. This paper proposes a framework for joint flood control operation among the identified key reservoirs in a transboundary river basin. We introduce a novel indicator, the Flood Control Contribution (FCC) indicator to identify key reservoirs among a large number of reservoirs in the basin. Taking the Lancang-Mekong River Basin (LMRB), a transboundary basin in Southeast Asia, as a case study, the flood control effectiveness of joint optimal operation for key reservoirs is investigated. Results show that the FCC-based key reservoirs are more recommended than the volume-based ones, which identifies nine of 92 reservoirs that contributed to half of the entire basin's flood control at lower coordination costs due to the involvement of fewer reservoirs. With the threshold of 90% of the total contribution, 26 reservoirs are identified as key reservoirs and jointly operated, yielding a 22-35% attenuation in the flood peaks of frequent floods and mitigating more than 78% excess flood volume of extreme floods. Notably, in the joint action, the upstream stakeholders (Laos and China) play a dominant role (controlling nearly 90% of the flooding water) in basin-wide flood control, which can provide significant flood control effectiveness to other downstream countries. The spatially heterogeneous flood control role and effectiveness of riparian countries present in transboundary joint action. This paper offers insights into basin system management and recommends policies to enhance flood control in the transboundary river.

Developing reliable hydrological models in highly managed basins is challenging due to multiple sources of uncertainty. The advent of open-source platforms providing publicly available datasets has the potential to mitigate these uncertainties. However, a comprehensive understanding of how these datasets impact model performance is lacking. This study takes the lower part of the YongDing River Basin (LYDRB) in northern China as a case to develop a hydrological model leveraging various open-source datasets, including water withdrawal activities, satellite-based streamflow, and remotely sensed evaporation. We design four comparative experiments to assess the impact of utilizing different data combinations on model performance. We find that the satellite- based streamflow data has the most significant impact, greatly enhancing streamflow simulation performance, with the NSE improving from the range of-1.5 to-0.39 to the range of 0.48 to 0.54 and the PBIAS improving from the range of-28 % to-63 % to the range of-3 % to-10 %. Water withdrawal data and remotely sensed evaporation data contribute to smaller performance improvements. The use of these two datasets may lead to poorer performance during the calibration period but better performance during the validation period. Specifically, remotely sensed evaporation data enhances model performance in streamflow simulation during the validation period, with NSE increasing by up to 0.1, although it results in a decrease of up to 0.04 in NSE during the calibration period. Overall, this study provides valuable insights for developing reliable and low-uncertainty hydrological models in highly managed and data-scarce basins by effectively utilizing various information sources.

Floodwater utilization is critical for balancing flood control and water conservation in multi-reservoir cascade systems, yet conventional reservoir operations often overlook the potential of storage substitution among different reservoirs. This study develops an extended flood control storage substitution model that accounts for both upstream-downstream interactions and intermediate reservoir regulation, thereby enabling a more precise assessment of flood control tasks substitution across multiple reservoirs. Focusing on the Jinsha River cascade and the Three Gorges Project (TGP), we systematically evaluate how additional flood control storage can be substituted and how to reserve storage to maximize hydropower benefits while maintaining flood safety. Results show that upstream reservoirs can collectively substitute as much as 3.3 billion m3 of TGP's design flood control storage, raising TGP's operating water level by up to 8 m and boosting system-wide power generation by 2.8 billion kWh. Moreover, strategic initial storage reservation-particularly assigning smaller additional storage to Baihetan and larger to Xiluodu-further enhances hydropower output without increasing flood risk. These findings underscore the importance of coordinated multi-reservoir operations that harness the "substitution effect" to optimize water concervation and electricity generation. By offering a generalizable model for quantifying reservoir substitution and guiding initial storage reservation decisions, this study provides valuable insights into floodwater resource utilization, promoting strategies to optimize water conservation, improve hydropower generation in multiple reservoirs systems.

Lakes in the Northeast Plains-Mountain Lake Region (NPLR) of China face severe risks of eutrophication due to climate change and intensive anthropogenic pressures. As a vital indicator for eutrophication status, the dynamics of chlorophyll-a (Chl-a) concentrations in NPLR lakes were, for the first time, comprehensively investigated in this study. A support vector regression (SVR)-based model was established and applied to the MERIS (2003–2011) and OLCI (2016–2019) observations to derive a long-term Chl-a record for 33 NPLR lakes. The NPLR lakes exhibited a climatological annual mean Chl-a of 12.3 mg m−3, ranging from 6.8 to 18.6 mg m−3 among the 33 studied lakes. During the study period, 11 lakes exhibited statistically significant increases in Chl-a concentrations (p < 0.05), and 3 lakes showed significantly decreasing trends. Spatially, lakes in lowland regions had higher Chl-a than those in mountainous regions. This study quantified the relative importance of seven explanatory factors in influencing interannual Chl-a changes for each lake. Results showed statistically significant impacts of agricultural fertilizer (10 lakes), wastewater (4 lakes), runoff (7 lakes), and wind (5 lakes) in affecting the interannual variations of Chl-a. The decreases in Chl-a were primarily attributed to the reduced livestock excrement. Compared to hydro-climatic factors, anthropogenic pressures (i.e. agriculture fertilizer, livestock excrement, and wastewater discharge) had more significant impacts on the interannual variations of Chl-a, accounting for more than 50% of 18 lakes. This study enhances our understanding of the long-term Chl-a dynamics in NPLR lakes and their responses to hydro-climatic factors and anthropogenic forcing. These findings are valuable for basin-scale water environment protection and sustainable development.

Upstream reservoir impoundment has brought adverse impacts on the downstream river-connected lakes, including lower water levels, frequent droughts, and reduced biodiversity. Existing research on ecological reservoir operation focused more on the effectiveness of operation redesign for ecology improvement, but less discussion on why they are eco-friendly. In this paper, the reservoir operation mechanisms targeting lake ecology are explored from three perspectives: ecological impoundment strategy, dominant characteristics of ecology and conservation tradeoffs, and critical parameters of the rule curves. Therefore, a reservoir operation framework, which couples the lake hydrological response model and multi-objective optimization model, is developed in this study. This framework is to maximize the water conservation and lake ecology benefits, assessing the latter by ecological reliability and ecological water shortage rate. Taking the Three Gorges Reservoir and Dongting Lake as the case study, operation rule curves are optimized for different impoundment strategies. The results indicate that impoundment with early storage time and slow storage rate can improve the ecological reliability and ecological water shortage by 9.90 % and 16.39 % than conventional operation. The ecological objectives are weakly competitive or synergistic when impounding fast, but strong competitive when impounding slow. Meanwhile, as the storage rate declines, the ecological improvement causes a slight decrease in power generation and a significant decrease in water supply. Finally, the crucial parameters of the rule curves are explored and appropriate values are recommended for different ecological preferences.

The complex and varied climatic conditions, short duration and high intensity of rainfall, and complex subsurface properties of semi-humid and semi-arid watersheds pose challenges for sub-daily flood forecasting. Previous studies have revealed that lumped models are insufficient because they do not effectively account for the spatial variability in hydrological processes. Extending the lumped model to a distributed modeling framework is a reliable approach for runoff simulation purposes. However, existing distributed models do not adequately characterize the high spatiotemporal variability in sub-daily hydrological processes. To address the above concerns, a distributed modeling framework was proposed that is extended from a lumped model and accounts for the effects of time-varying rainfall intensity and reservoir regulation on hydrological processes. The results indicated that the proposed distributed model could simulate sub-daily flood events with mean values of the NSE, BIAS, RPE, and PTE evaluation metrics of 0.80, 9.2

Reservoir operation is an effective measure for flood disaster reduction. Previous studies focused on minimizing flood risk to reduce potential flood damage, however, failure and recovery processes are not considered. To address this shortcoming, this study introduces the concept of resilience to measure the ability of flood control systems to maintain or return to the original states after an extreme flood. The quantitative expression of resilience, resilience metric, is proposed based on the system performance and defined as the summation of system functionality loss throughout the entire flood process. Based on the proposed resilience metric, a multiobjective optimization model with flood risk represented by maximum reservoir water level and downstream peak flow, and flood resilience as objectives, is established. Nierji Reservoir, located in Northeast China is taken as a case study. Results show that the proposed method can improve resilience without increasing flood risks (i. e., maximum reservoir water level and downstream peak flow) compared with the traditional model. Pareto optimates of the optimization model show that there exist tradeoffs between risk and resilience, improving resilience inevitably reducing risk. The resilience improvement is large when the flood risk of the reservoir or downstream is medium and is limited when the flood risk is high. Moreover, the resilience improvement is higher for a smaller flood. The improvement is achieved through pre-releasing at the early stage of the flood and releasing slowly at the recession stage of the flood. Six typical schemes, selected referring to the existing scheme are comprehensively compared, and an operation rule for flood control with the highest resilience but without reducing risks is recommended. This study provides a new way for flood control management.

Utilizing the spatial heterogeneity and climate periodicity of various available renewable energy sources can enhance the multienergy complementarity, which will further reduce the energy storage demand and contributes to the "virtual energy storage gain." In this paper, we propose a spatiotemporal coordination method based on spectral analysis for a wind-PV-hydropower system that targets the maximum virtual energy storage gain. The complementary effect of hydropower on wind and PV power can be seen as changes in the regulation ability resulting from the hydropower construction development as well as a decreased variance in the total system production output. This method is used to determine the optimal coordination distance of multiple energy sources that are matched over different periods representing typical climate variation. A case study in the Yalong River basin in China obtains the best range of hydropower bundling surrounding wind power and PV power stations under different hydropower construction planning scenarios, and it reveals that increased regulation capacity of hydropower will reduce the optimal coordination distance but can achieve more energy balance if both actual storage and virtual storage is accounted for. This approach can realize energy delivery with the optimal coordination distance to meet intensive and efficient development needs, which can provide guidance and support for the planning and construction of wind-PV-hydropower storage systems.

Water use was impacted significantly by the COVID-19 pandemic. Although previous studies quantitatively investigated the effects of COVID-19 on water use, the relationship between water-use variation and COVID-19 dynamics (i.e., the spatial-temporal characteristics of COVID-19 cases) has received less attention. This study developed a two-step methodology to unravel the impact of COVID-19 pandemic dynamics on water-use variation. First, using a water-use prediction model, the water-use change percentage (WUCP) indicator, which was calculated as the relative difference between modeled and observed water use, i.e., water-use variation, was used to quantify the COVID-19 effects on water use. Second, two indicators, i.e., the number of existing confirmed cases (NECC) and the spatial risk index (SRI), were applied to characterize pandemic dynamics, and the quantitative relationship between WUCP and pandemic dynamics was examined by means of regression analysis. We collected and analyzed 6-year commercial water-use data from smart meters of Zhongshan District in Dalian City, Northeast China. The results indicate that commercial water use decreased significantly, with an average WUCP of 59.4%, 54.4%, and 45.7%during the three pandemic waves, respectively, in Dalian. Regression analysis showed that there was a positive linear relationship between water-use changes (i.e., WUCP) and pandemic dynamics (i.e., NECC and SRI). Both the number of COVID-19 cases and their spatial distribution impacted commercial water use, and the effects were weakened by restriction strategy improvement, and the accumulation of experience and knowledge about COVID-19. This study provides an in-depth understanding of the impact of COVID-19 dynamics on commercial water use. The results can be used to help predict water demand under during future pandemic periods or other types of natural and human-made disturbance.

The merging of multi-source precipitation retrievals (PRs) and gauge-based observations (GO) provides a new opportunity for precipitation field estimation. However, the uncertainties associated with PRs may become relatively more evident when the gauge network better captures the spatial distribution of the rainfall fields. To dynamically balance the utilisation of PRs, we propose a merging framework based on a novel concept; namely, the virtual gauge, where the grid cells utilising PRs are regarded as virtual gauges, and the framework is henceforth referred to as VG. The main steps were as follows: i) determine the locations of virtual gauges from multi-source PRs, ii) estimate rainfall at virtual gauges using a basic merging method, and iii) spatially inter-polate using both actual and virtual gauges. Accordingly, the case study employed random forest and inverse distance weight as the basic merging and interpolation methods of VG. Evaluation using real world data over a region where nearly each 0.1 degrees grid cell contains a ground gauge indicates that VG improves around 7-11% over its basic methods and improves around 32-240% over its inputting PRs. VG performed better and was more stable than the basic methods under various gauge densities, rainfall intensities, and rainfall distributions. The results showed that VG could supplement the spatial information of rainfall fields missed by the gauge network and reduce interference caused by the uncertainties of PRs. Overall, the framework integrates the advantages of existing merging and spatial interpolation methods by adjusting the number, locations, and rainfall estimations of virtual gauges.

Joint operation of the reservoir system can greatly improve flood water utilization and alleviate the contradiction between water supply and water demand. But yet, the flood risk brought by forecast uncertainties transfers between the reservoirs and propagates from upstream to downstream, which is crucial for flood water utilization but is rarely investigated. To address this issue, this study proposes a FLWL optimal control framework, which consists of three modules: the FLWL upper bound derivation module, the flood risk analysis module, and the optimal FLWL decision-making module. The first module is to derive the upper bound relationship of the flood limited water level (FLWL), i.e., the specific substitution relationship of storage, for the parallel reservoirs coupling the capacity-constrained pre-release method and the aggregation-decomposition method. In the second module, the method to calculate the flood risk of the reservoirs and downstream caused by rainfall and runoff forecast uncertainties based on the total probability formula is proposed. In the third module, the transmission and propagation of flood risks in the parallel reservoir system under different rainfall grades are quantified, based on which the optimal FLWLs are determined. The implication in the Qinghe-Chaihe parallel reservoir system located in Northeast China shows that FLWL in Qinghe Reservoir and Chaihe Reservoir can be raised to 129.5 m from 127 m and 108 m from 104 m, respectively. Controlling the FLWL at any point in the safety domain can improve both flood water utilization and flood control benefits. However, the flood risk nonlinearly transfers between the parallel reservoirs and nonlinearly propagates from the upstream to the downstream. Based on this relationship, the optimal FLWLs of the reservoirs under different decision preferences are recommended.