Flash drought (FD) has become a major risk to summer maize production and agricultural water management in the Huang–Huai–Hai (HHH) Plain. However, previous studies have mainly focused on end-of-season yield losses or changes in event frequency, with limited causal evidence on how FDs disrupt crop growth processes within the growing season, particularly across phenological stages. In this study, we examined the length of growing season (LOS), vegetative growth stage (VGS), and reproductive growth stage (RGS) to identify the average treatment effects of FD occurrence, the average marginal effects of increasing FD intensity, and their spatiotemporally heterogeneous responses. Using multi-source data from 2001 to 2020, we combined Double Machine Learning and causal forests to estimate occurrence and intensity effects, and used Gross Primary Productivity (GPP)-based time-frequency analysis for process-level corroboration. FD occurrence showed significant average treatment effects on all three phenological indicators. Under the continuous treatment pathway, significant average marginal effects were detected for VGS and RGS, whereas the effect on LOS was not significant. Irrigation water use, temperature, and precipitation were identified as key regulators of effect heterogeneity. Moreover, FD impacts were not temporally stationary, but were continuously restructured during 2001–2020. Time-frequency coupling analysis of GPP further indicated that FD risk reflects the combined effects of crop growth dynamics, hydrothermal conditions, and irrigation management. These findings shift the assessment of FD impacts from end-of-season losses to in-season growth-process perturbations, and provide quantitative support for sub-seasonal agricultural drought early warning and stage-specific irrigation scheduling.

Highlights What are the main findings? What are the implications of the main findings?Highlights What are the main findings? What are the implications of the main findings?Abstract Forest-steppe ecotones exhibit pronounced spatiotemporal heterogeneity and complex climate-vegetation interactions, posing significant challenges for vegetation dynamics prediction. Existing models often struggle to capture long-range temporal dependencies, preserve spatial continuity across heterogeneous transition zones, and provide ecologically interpretable insights. To address these limitations, we developed a bidirectional Geo-Spatial Mamba (Geo-S-Mamba) architecture with a multi-objective loss function incorporating spatial continuity constraints based on the first law of geography. The model was trained using multi-source geospatial datasets and independently validated during 2019-2023. The results show that Geo-S-Mamba achieved an R 2 of 0.93. Moreover, both the bidirectional mechanism and the spatial-continuity loss improved the PSDI by approximately 0.08. The model effectively captured annual variations in NDVI and covariation among vegetation groups. Post hoc symmetric causal learning based on Pearl's structural causal theory indicated that precipitation was the primary driver of grassland vegetation dynamics. Temperature and radiation influenced NDVI mainly through boundary-dependent effects. Overall, this framework can estimate changes in the spatial distribution of plant communities across heterogeneous environments and provides a scientific basis for further research on forest-steppe ecotones.

To enhance the flood simulation accuracy of the Xin'anjiang (XAJ) model, an improved Xin'anjiang model (I-XAJ) was developed by incorporating an aggregated reservoir module with storage and discharge functions into the original framework. Remote sensing technology was employed to extract information on water storage structures within the basin. In the aggregated reservoir module, a storage capacity curve was established to analyze the retention effects caused by the non-uniform spatial distribution of small water storage bodies, and outflow was calculated using the weir-flow formula for ungated spillways. To evaluate the forecasting accuracy of the I-XAJ model, thirty-one flood events were selected from the Huangnizhuang (HNZ) basin. Compared with the XAJ model, simulation accuracy of floods was significantly improved by the IXAJ model through consideration of small water bodies within the basin. The model performance for flood peaks was increased by 44.5 percentage points in the calibration period and 25 percentage points in the validation period.

Climate warming is intensifying drought stress on terrestrial carbon uptake, particularly in climate transition basins where vegetation productivity depends on atmospheric demand and antecedent water supply. However, drought studies based on gross primary productivity (GPP) often treat reconstruction, attribution, threshold analysis, compound-stress diagnosis, and recovery as separate tasks, which limit hydrological interpretation. Here, monthly 1-km GPP anomalies and hydroclimatic predictors from 2001 to 2024 were used to examine how atmospheric evaporative demand and antecedent hydrological memory regulate productivity loss, compound drought amplification, and post-drought recovery in the Huai River Basin. A PSO-LSTM-Transformer reconstruction generated monthly GPP, which were then analyzed using SHapley Additive exPlanations, a Synergy Index, and Time-to-Recovery. Attribution identified vapor pressure deficit (VPD), five-month antecedent precipitation (Ante_prec_5), and three-month antecedent temperature as the dominant controls. VPD and Ante_prec_5 represented the demand and supply sides of drought response, with basin-specific transition points near 0.82 kPa and 233 mm. About 11.06% of the basin showed non-additive GPP loss under compound atmospheric-hydrological drought. Event-scale analysis indicated that antecedent hydrological deficits reduced buffering capacity before peaking atmospheric dryness, after which limitation shifted toward atmospheric-demand stress. Recovery was faster in the humid south and slower in the northern plains, and Time-to-Recovery increased by about 0.20 months per degree latitude. Ante_prec_5 constrained recovery more strongly than VPD, indicating that rebuilding hydrological supply is central to post-drought productivity recovery. This demand-supply framing supports drought monitoring, ecological water availability assessment, and resilience evaluation in water-sensitive transition basins.

To address the limitations of existing groundwater quality evaluation methods, such as subjective weighting, ambiguous classification boundaries, and the lack of multi-perspective robustness verification, this study constructed two quantitative groundwater evaluation frameworks by coupling the Projection Pursuit (PP) model with the Set Pair Analysis (SPA) model. First, a PP-based evaluation model was established. An Accelerated Genetic Algorithm (AGA) was employed to optimize both the projection direction and the parameters of the Logistic curve, thereby fully exploiting the intrinsic structural features of high-dimensional water quality data to achieve objective dimensionality reduction and continuous grade quantification driven by multi-index data. Second, to overcome the ambiguity at the classification boundaries of water quality grades and to validate the evaluation results, the SPA model was further introduced for fuzzy evaluation. Innovatively, the optimal projection direction derived from the PP model was utilized as the objective weight in the SPA model to quantify the degree of membership of the samples to different water quality grades. Taking the groundwater in the main urban area of Bozhou City, Anhui Province, as the study subject, the results indicated that the overall groundwater quality in this area belongs to Grade II (Class II water). The characteristic values of the water quality grades calculated by the PP model ranged from 1.89 to 2.00, while those calculated by the SPA model ranged from 1.95 to 2.03. The evaluation results of these two models, which are based on different mathematical mechanisms, exhibited high consistency. Furthermore, index perturbation sensitivity analysis and independent weight control experiments systematically verified the robustness of the evaluation results and the scientific validity of the weighting method. This study provides a rigorous theoretical basis and technical support for the precise evaluation and sustainable management of regional groundwater.

The traditional stability assumption of dependence structure constructed by diverse hydro-meteorological variables has been altered, and the non-stationarity characteristic and its driving factors recognition of dependency structures are crucial for drought disaster prevention analysis. In this study, we utilized the copula-based likelihood ratio (CLR) method to reveal the non-stationarity features of the dependency structure between precipitation (P) and its influencing factors including temperature (T), relative humidity (RH) and soil moisture (SM). Also, the driving factors of dependency structures P-T, P-RH and P-SM series were illustrated using an Random Forest (RF) approach. The application results in Anhui Province, China, indicated: (1) evident non-stationarity features were identified for the evolution of the P-RH series in seven cities; (2) SM, T, SD and ENSO are the primary driving factors both before and after the variability points of P-RH series. These findings will be favourable for understanding the occurrence mechanism of non-stationarity characteristics of hydrological variables.

Study region: The Ganjiang River Basin, a major humid subtropical tributary of the Yangtze River located in southeastern China. Study focus: This study develops a lightweight and interpretable precipitation merging framework based on Absolute Distance Inverse Weighting (ADIW) to integrate eight mainstream precipitation datasets (CHIRPS, CMORPH, ERA5, GSMaP, IMERG, PERSIANN, SM2RAIN, TRMM) from 2008 to 2020. The performance of the merged product, corrected with four methods (Linear Regression-LR, Linear Scaling, Quantile Mapping, Quantile-Quantile), was rigorously evaluated through hydrological simulation using the HYPE and VIC models. New hydrological insights for the region: Results show that: (1) The ADIW merging framework effectively synthesizes multiple datasets, yielding a product with superior correlation and rainfall detection skill compared to any individual input, despite a slight underestimation. (2) While all four bias-correction methods effectively mitigated systematic errors, Linear Regression (LR) proved the most robust and consistent in enhancing both precipitation accuracy and subsequent runoff simulations. (3) Consequently, the combined ADIW+LR approach delivered the optimal hydrological performance, achieving the highest Nash-Sutcliffe and Kling-Gupta efficiency values in both models. (4) Diagnostic analysis identified ERA5 and GSMaP as the most influential datasets contributing to the merged product, and established that relative bias (RB) and mean absolute error (MAE) are the key metrics controlling hydrological reliability.

Crop and coupled hydrological crop models are used to simulate the hydrological processes on agricultural land. However, these models tend to oversimplify the hydrological and groundwater modules, respectively, resulting in less accurate simulations of macroscopic hydrological processes. To address this issue, the Variable Infiltration Capacity-Environmental Policy Integrated Climate (VIC-EPIC) model was used as a basis for introducing the groundwater module (HYDRUS) and a scheme was proposed to parameterize the effects of irrigation on crop evapotranspiration. A coupled VIC-EPIC-HYDRUS (VEH) model was developed, and its simulation accuracy was validated. The validation results for the Qingkou River Basin in northern Jiangsu indicated that the VEH model achieved superior performance with 17.1%, 4.1%, and 14.5% lower total errors compared to VIC-HYDRUS, VIC-EPIC, and Soil & Water Assessment Tool (SWAT) models, respectively. Among the integrated modules, the EPIC module contributed most significantly to model improvement, while the HYDRUS module specifically enhanced soil moisture simulation accuracy by 20.4%. Comparative analysis of different calibration approaches showed that incorporating remote sensing evapotranspiration data decreased evapotranspiration error slightly, whereas using ground-observed soil moisture data reduced total error by 29.0%, demonstrating the importance of high-quality observational data for model calibration. Incorporating crop, groundwater modules, and high-quality observational data into a hydrological model substantially improves the simulation accuracy of hydrological elements, making it highly applicable for promotion. These findings can help policymakers make informed decisions regarding water management practices.

Under global climate change, precipitation variability and extreme events pose significant challenges to regional ecological security and water resource management. This study proposes the Dynamic Line Scanning Method (DLSM) to quantitatively assess the migration of China's 800 mm precipitation isohyet during 1961-2022. Using long-term precipitation data, we systematically examined its spatial and temporal variation, explored links with drought-flood regimes and extreme precipitation, and identified the main drivers. Results indicate that the isohyet experienced a two-phase shift: an initial southward retreat followed by accelerated northward movement after 2001, with the latter trend markedly intensifying. These shifts have directly influenced regional drought-flood patterns and altered the frequency and intensity of extreme precipitation events. Analysis further reveals that the East Asian Summer Monsoon Index (EASMI) is the dominant factor driving the isohyet's movement. Overall, this study provides novel methodological insights and robust empirical evidence regarding the dynamics of precipitation isohyets in China in the context of climate change. The findings enhance understanding of hydroclimatic variability and offer a scientific foundation for developing region-specific adaptation and water management strategies.

Multivariate drought prediction analysis is of great significance for drought disaster resistance. In this study, based on the derivation of monthly time-variant drought indicators SPIt, STIt and SSMIt by means of GAMLSS method, the conditional variables and their optimal combination structure were recognized through correlation analysis between targeted agricultural drought prediction variable SSMIti-1 and previous 1–12 monthly drought indicators SPIti-1-n, STIti-1-n and SSMIti-1-n, then the response relationship between combined distribution of conditional variable under different patterns and occurring probability of agricultural drought events was revealed through C-Vine Copulas function and conditional quantile regression method, and finally the CQRM approach based on C-Vine Copulas function of agricultural drought prediction analysis was proposed, which was verified through its application in northern Anhui Province area, China. It can be summarized from the application results that, (1) generally, previous 1–3 monthly-scale drought indicators SPIti-1-n, STIti-1-n and SSMIti-1-n can be utilized as primary conditional variables of agricultural drought prediction analysis, and the response time of drought propagation process to diverse conditional variables in summer and autumn (less than 1 month) is relatively shorter in comparative with that of winter and spring (nearly 3 months). (2) the utilization of more conditional variables is not definitely beneficial for the enhancement of agricultural drought prediction accuracy, and it is crucial to recognize effective monthly conditional variables and also construct its optimal combination structure through correlation analysis to improve the overall fitting performances of agricultural drought prediction analysis.

Continuous droughts occur frequently worldwide, and drought damage sensitivity assessment is a core link in drought risk management. This study uses crop experiments in the Huaibei Plain to simulate maize growth processes under various continuous drought by a localized AquaCrop model. Then, a crop transpiration drought index (CTDI) is proposed, and three-phase drought damage sensitivity curves between the drought intensity (DHI) and yield loss rate are established. Thereupon, the drought damage sensitivity of maize yield formation is assessed. The results show that the CTDI considers the actual crop water supply-demand balance and growth response under continuous drought, thereby it can accurately identify the real agricultural drought status. Furthermore, slight deficit irrigation can be carried out at the seedling stage, and moderate deficit irrigation can be implemented at the milking stage for maize. Nevertheless, severe drought during the seedling stage should be avoided to ensure the survival of plants. Moreover, when plants have suffered from drought during the seedling stage, an adequate water supply should be guaranteed at the jointing stage. Therefore, drought damage sensitivity curves have practical implications for rapid drought loss estimation and decision-making for drought reduction strategies in the Huaibei Plain. [GRAPHICS]

The coordinated development of water resources, social economy, and ecological environment (WSE) in the irrigation district is fundamental to sustainable management and food security. In this study, a model is established to address the lack of effective methods for quantitative assessment and diagnosis of WSE coordinated development due to the complex and uncertain WSE structure. Therefore, an assessment model of WSE development is proposed by utilizing the structural analysis of WSE and connection number theory. This study adopts a mechanical model to assess the status of WSE coordination and diagnose the uncoordinated reasons, and a coordinated development assessment model is constructed using a risk matrix. Applying this method in five counties of the Dagong irrigation district of the Yellow River indicates average values of 0.403 (low development), 0.377 (medium slowness), 0.497 (low development), 0.572 (medium development), and 0.589 (medium development) for the WSE development degree in Neihuang, Xunxian, Huaxian, Changyuan, and Fengqiu, respectively, during 2010–2017, with average values of 0.677 (medium incoordination), 0.781 (low coordination), 0.611 (high incoordination), 0.797 (low coordination), and 0.842 (medium coordination) for the WSE coordination degree. Coordinated development status in the area deteriorates gradually from south to north, and although the overall situation has been improved, the differences among the five counties remain significant, with Neihuang and Huaxian being the most severe incoordination. Enhancing the advance of economy and environment, matching and promoting the support capacity of water resources condition, and maintaining a balance of these three sub-systems are the crucial points of coordinated development of the WSE structure in the Dagong irrigation district. This study provides targeted policy reference for water resources planning and effectively promotes the application of structural water resources science.

Traditional coupling coordination degree (CCD) evaluation methods fail to simultaneously ensure the accuracy and reliability of evaluation results. To overcome this limitation, a novel evaluation method that integrates the logical multiplication of connection numbers with stochastic simulations (ECCD-LMS) is developed to assess the CCD of regional water resources, social economy, and ecological environment (WSE) composite systems. The method adopts three-element connection numbers to quantify the comprehensive evaluation level of each system. Overall partial connection numbers and triangular fuzzy numbers then define dynamic value intervals for the connection number components, and the Monte Carlo method is integrated to simulate stochastic variation in each component. The simulated values are substituted into the logical multiplication of connection numbers to generate evaluation results that include both mean CCD estimates and 95% uncertainty intervals. The empirical application of ECCD-LMS in China’s Jing River Basin indicated that the coupling coordination level of the regional WSE composite system exhibited an overall increasing tendency with fluctuations from 2012 to 2023. These fluctuations were closely associated with variations in the comprehensive evaluation level of the water resources system. Spatially, the disparities in evaluation grades among subregions gradually narrowed during the study period. Compared with traditional CCD evaluation methods, ECCD-LMS effectively corrects systematic overestimation while maintaining objectivity and produces more dispersed CCD estimates that reflect differences among evaluation samples with greater clarity. Furthermore, by outputting mean CCD estimates and 95% uncertainty intervals, ECCD-LMS outperforms traditional methods that only provide static point estimates. It thus enables robust and credible evaluation of the coupling coordination level of WSE composite systems under multiple sources of uncertainty, suggesting potential applicability across diverse regional contexts.

Precipitation plays a vital role in the hydrological cycle, directly affecting water resource management and influencing flood and drought risk prediction. This study proposes a Bayesian Model Averaging (BMA) framework to integrate multiple precipitation datasets. The framework enhances estimation accuracy for hydrological simulations. The BMA framework synthesizes four precipitation products—Climate Hazards Group Infrared Precipitation with Station (CHIRPS), the fifth-generation ECMWF Atmospheric Reanalysis (ERA5), Global Satellite Mapping of Precipitation (GSMaP), and Integrated Multi-satellitE Retrievals (IMERG)—over China’s Ganjiang River Basin from 2008 to 2020. We evaluated the merged dataset’s performance against its constituent datasets and the Multi-Source Weighted-Ensemble Precipitation (MSWEP) at daily, monthly, and seasonal scales. Evaluation metrics included the correlation coefficient (CC), root mean square error (RMSE), and Kling–Gupta efficiency (KGE). The Variable Infiltration Capacity (VIC) hydrological model was further applied to assess how these datasets affect runoff simulations. The results indicate that the BMA-merged dataset substantially improves precipitation estimation accuracy when compared with individual inputs. The merged product achieved optimal daily performance (CC = 0.72, KGE = 0.70) and showed superior seasonal skill, notably reducing biases in autumn and winter. In hydrological applications, the BMA-driven VIC model effectively replicated observed runoff patterns, demonstrating its efficacy for regional long-term predictions. This study highlights BMA’s potential for optimizing hydrological model inputs, providing critical insights for sustainable water management and risk reduction in complex basins.

The quantitative assessment of regional agricultural drought risk is of great importance for the mitigation of regional droughts and the maintenance of food security. However, the complex structure of the drought risk system and the inherent uncertainty result in a lack of effective assessment methods. The study constructed a comprehensive assessment index system based on the six-element drought risk system. A novel Mutually Exclusive Relative Connection Number synthesizing the Fuzzy Risk Matrix (MERCN-FRM) method was proposed to assess agricultural drought risk. Additionally, a CRITIC method was improved through the integration of the set pair potential and entropy to weight the index. A case study was conducted on the Jianghuai watershed area located in the middle of east China with frequent drought disasters. The results demonstrate that the six-component drought risk framework offers a holistic view of how the risk factors interact. The enhanced CRITIC method can comprehensively represent the positive, negative, and uncertain fuzzy correlations of indicators microscopically. The MERCN-FRM methodology, through the utilization of MERCN, achieves a dynamic synthesis of the pessimistic and optimistic risk matrices. This approach not only surmounts the subjectivity and static nature inherent in traditional risk matrix synthesis methods, but also facilitates a more profound understanding of the interactive mechanisms within the subsystems of drought risk compared to other widely-used comprehensive evaluation methods. This approach could be employed as a novel tool for dynamically assessing regional disaster risk, aiding the decision-making for drought risk mitigation strategies.

The implement of water resources spatial equilibrium (WRSE) schemes is fundamental task of integrated water resources management in China, in which, the evaluation and simulation analysis of WRSE system is of great concern for understanding the overall variation and feedback characteristics of WRSE system. Therefore, we utilized the ordered degree, entropy and connection number coupling model to evaluate the variation of WRSE system, and also employed system dynamics (SD) and scenario simulation integrated method to reveal the feedback characteristics between different equilibrium variables and subsystems, and thus the set pair analysis-SD based approach for structure simulation and variation trend evaluation of WRSE system was constructed. The application results in Anhui province, China demonstrated that, the overall variation of provincial spatial equilibrium situation of WRSE system presented obvious improving trend, 2009-2019, the index of ordered degree and connection number entropy in Hefei, increased from the minimum of 0.6434 (Grade 3, unequilibrium) to maximum of 0.9985 (Grade 1, equilibrium). Moreover, the future variation of WRSE system will display stable improving trend during 2020 to 2029, in which, annual water resources availability and water resources utilization efficiency will have significant influence on WRSE system. And the research findings can be favorable to formulate water resources development and utilization strategies.

Flood forecasting is regarded as the most important basic non-engineering measure, and its accuracy is the key to scientific flood control and regulation. The conceptual rainfall–runoff model (CRR) is widely applied to flood forecasting. The major difficulty associated with the use of CRR models in hydrology is their calibration since most of these models involve a large number of parameters. In order to calibrate the parameters of the CRR model, an improved quadratic interpolation optimization algorithm (IQIO) was proposed. The tent chaos mapping was used to initialize the population, adaptive optimizer probability based on individual adaptation value was used to balance algorithm’s global exploration and local exploitation ability. Thirteen mathematical benchmark functions were used to test the IQIO algorithm. The results showed that the IQIO algorithm exhibited strong exploration capability and fast convergence speed. The CRR model parameters optimized by the IQIO algorithm exhibited high performance, with Nash–Sutcliffe efficiency (NSE) values reaching 0.951 during the calibration period and 0.913 during the validation period. The relative error of runoff in each year was less than 20%, which satisfied the calculation accuracy requirements.

In order to deeply reveal the application characteristics of set pair analysis in structural water resources methodology and look forward to its main development trends,combined with existing research results,based on the methodological perspective of relational structure,a three-element structural principle of structural water resources methodology system consisting of research object,research variable,and research objective(abbreviated as the Ternary Structure Principle)was summarized and proposed.Based on the principle,the application progress of ensemble analysis in modeling,optimization,identification,simulation,prediction,evaluation,decision-making,reasoning and other methods of structural water resources methodology was systematically expounded.It is pointed out that set pair analysis had obvious advantages in describing and handling the uncertain argumentation relationship between research variables and research objectives.The implementation process of the first seven methods reflects the information transmission between the research object,research variables,and research objectives,while the implementation process of the inference method reflects the information transmission between the evidence,argumentation,and conclusion relationship.Compared with the first seven methods,the abstraction and complexity level of the water resources system inference method are higher.The research approach of combining physical cause analysis of uncertainty formation process with quantitative calculation is an important development trend in the future to explore more systematic and pioneering water resource set pair analysis methods.

The development philosophy of water resources spatial equilibrium (WRSE) is a crucial aspect of China's water conservancy strategy. To quantitatively evaluate the WRSE in Chongqing and diagnose obstacle indicators, the WRSE evaluation and diagnostic model was proposed based on coupling coordination degree and subtraction set pair potential (SSPP) from both water supply and demand perspectives. The results showed that economically developed regions in Chongqing were suffering water scarcity and disequilibrium. The WRSE state from excellent to poor was the city cluster of Wuling mountain area in southeastern Chongqing (CSC), the city cluster of the Three Gorges Reservoir area in northeastern Chongqing (CNC), the new area of Chongqing city proper (NAC), and the central urban area of Chongqing (CAC). The obstacle indicators in CAC, NAC, CNC, CSC, districts, and counties were diagnosed by an improved diagnostic method based on SSPP, which can avoid the unreasonable diagnostic result. The evaluation and diagnostic results at different spatial scales can provide a more comprehensive reference for water resources management, and the results are consistent with the actual conditions of Chongqing. Our study can provide insights for WRSE evaluation and the diagnosis of obstacle indicators. It also presents a method that can be applied to other systems, including environmental and resource management, across various spatial scales.