Water Resources Research最新文献

{"title":"Interpretable Soil Water Retention Prediction Using Hierarchical Attention Networks With Uncertainty Quantification","authors":"Teamrat Afewerki Ghezzehei","doi":"10.1029/2025wr042833","DOIUrl":"https://doi.org/10.1029/2025wr042833","url":null,"abstract":"Abstract Understanding which soil properties control water retention at different soil moisture states remains critical for predicting soil responses to management and environmental change. Traditional pedotransfer functions predict parameters of equations like van Genuchten's model, but parameter correlations create ill‐posed inverse problems that propagate errors. Fundamentally, parameter‐based approaches cannot reveal which properties matter most at different water potentials or how texture‐structure interactions vary across the moisture spectrum. This study presents Hierarchical Attention‐Based Pedotransfer Function (HABIT) to address these limitations through two objectives: (a) achieve superior prediction accuracy through direct soil moisture prediction that circumvents parameter correlation errors, and (b) quantify property importance and interactions across soil moisture regimes through interpretable attention mechanisms. The model employs attention architectures that dynamically weight soil properties based on water potential, enforces physical constraints including monotonicity, and accommodates incomplete soil characterization through hierarchical training. HABIT is evaluated against traditional approaches using comprehensive international soil databases spanning global textural and structural variability. Attention weight analysis reveals moisture‐dependent property interactions, including unequal bidirectional texture‐structure interactions where texture exerts stronger influence on structural interpretation than the reverse, organic carbon's dual structural and sorptive roles, and saturated hydraulic conductivity's role as a statistical surrogate encoding pore network information through correlations with unmeasured properties. These patterns generate testable predictions while explaining how direct prediction with adaptive property weighting captures soil‐specific retention behaviors that parametric forms miss. HABIT demonstrates that properly designed machine learning architectures serve as both prediction tools and hypothesis‐generating frameworks for quantitative analysis of soil property interactions.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"8 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148893523","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Global Transpiration Changes and Its Drivers in the Future","authors":"Jie Wu,Chunmiao Zheng,Zhenzhong Zeng,Guy Schurgers","doi":"10.1029/2025wr042783","DOIUrl":"https://doi.org/10.1029/2025wr042783","url":null,"abstract":"Abstract Plant transpiration ( E t ), the dominant element of evapotranspiration, plays an important role in hydrological, ecological, and climate processes. However, attributing projected future E t change to elevated atmospheric carbon dioxide concentration ([CO 2 ]), air temperature ( T a ), and leaf area index (LAI) remains challenging because these drivers exert opposing and interacting influences. Here, we develop a quantitative first‐order attribution framework that combines grid‐cell‐level sensitivities diagnosed from CMIP6 single‐forcing experiments with projected driver changes under Shared Socioeconomic Pathways (SSPs). Our results show that simulated future global E t shifts from +10.42 ± 11.84 mm yr −1 (+3.9%) under SSP1‐2.6 to −14.33 ± 29.91 mm yr −1 (−5.4%) under SSP5‐8.5, largely because the diagnosed negative physiological contribution of elevated [CO 2 ] increasingly exceeds the positive contributions associated with increasing T a and LAI under stronger forcing. Spatially, the [CO 2 ]‐related reduction is widespread, whereas the T a ‐related contribution is more pronounced in humid regions and the LAI‐related contribution is larger in arid and semi‐arid regions. Inter‐model spread in the attributed contributions arises substantially from differences in diagnosed sensitivities of Δ E t to changes in [CO 2 ], T a , and LAI, with the LAI‐related term showing the largest spread among the three first‐order terms considered. These results provide a consistent quantitative basis for interpreting how elevated [CO 2 ], warming, and vegetation dynamics jointly shape projected E t and highlight the need to better represent plant physiological responses, vegetation structural dynamics, and land‐atmosphere coupling in Earth System Models.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"20 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148932878","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Global Sensitivity Analysis of a Reactive Transport Model for Abiotic Trichloroethylene Reduction by Reduced Iron Minerals in Clay‐Rich Aquitards","authors":"G. Kumar,H. S. Lim,A. Störiko,A. J. Valocchi,C. J. Werth","doi":"10.1029/2026wr044244","DOIUrl":"https://doi.org/10.1029/2026wr044244","url":null,"abstract":"Abstract Back‐diffusion of trichloroethylene (TCE) from clay‐rich aquitards poses a persistent challenge to groundwater remediation at many contaminated sites. Previous studies have demonstrated that natural attenuation of TCE is controlled by abiotic reactions with reactive ferrous iron; however, the coupled roles of mineral inventory, mineral reactivity, and biogeochemical processes governing reduced iron formation remain poorly understood. In this study, a global sensitivity analysis was performed on a reactive transport model using the Distance‐based Generalized Sensitivity Analysis method to identify key parameters that control abiotic reduction of TCE in clay‐rich aquitards under anoxic conditions. Uncertain input parameters including sediment composition, surface area, intrinsic reactivity, thermodynamics, and microbial kinetics were varied across plausible ranges, and model realizations were clustered based on quantities of interest related to TCE degradation and mineral phase evolution. Results show that TCE attenuation is controlled by the initial inventory, reactive surface area of reduced iron minerals and by their surface‐area‐normalized intrinsic reaction rates with TCE. Across most realizations, TCE attenuation in clay‐rich aquitards can be effectively approximated as a first‐order decay process when sufficient reactive iron is present. Secondary controls associated with biological iron and sulfate reduction processes influence TCE degradation only under a narrow subset of conditions characterized by limited initial reduced iron availability, where mineral regeneration becomes rate‐limiting. These findings provide a quantitative basis for simplifying reactive transport representations of abiotic TCE degradation and highlight the conditions under which detailed biogeochemical coupling is necessary for reliable prediction of long‐term abiotic natural attenuation of TCE in clay‐rich aquitards.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"32 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148932881","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Camp Creek Revisited, One More Time: Evaluating Multiple Approaches to Quantify Peak Flow Response to Forestry for a Medium‐Size Snow‐Dominated Catchment","authors":"R. Dan Moore,Kim Green,Cydne Potter","doi":"10.1029/2026wr043696","DOIUrl":"https://doi.org/10.1029/2026wr043696","url":null,"abstract":"Abstract Forest disturbance effects on peak flows have been debated for decades. Paired‐catchment analysis (PCA) has been criticized in relation to the long‐term stability of pre‐harvest regressions and the inability of chronological pairing to characterize changes in event frequency. Non‐stationary frequency analysis (NFA) avoids some concerns with PCA. However, a fundamental problem is that the concept of return period is ambiguous under non‐stationary conditions. This study re‐examined peak flow response to forestry at Camp Creek, a snow‐dominated 34‐km 2 catchment in British Columbia, using five approaches: (a) chronologically paired PCA using adjacent Greata Creek as a control; (b) frequency‐paired PCA using maintenance‐of‐variance regression to fit the pre‐harvest relationship; (c) time‐series regression to relate Camp Creek annual maximum flows to annual peak snowpack water equivalent (SWE) and equivalent clearcut area (ECA); (d) NFA using SWE and ECA as covariates; (e) the design‐life‐level (DLL) approach. While all approaches reveal aspects of forestry effects on peak flows, only DLL provides a conceptually meaningful probabilistic approach to risk analysis under non‐stationary forest cover. However, the DLL‐based approach involves the application of NFA using ECA as a covariate. ECA may not provide a robust indicator of the effects of forest disturbance on peak flows in snow‐dominated catchments, particularly for situations in which logging occurs progressively over extended periods of time or where disturbance also occurs due to fire, insect infestation or disease. Further research should focus on developing alternative indices of forestry effects on peak flow, and on testing process‐based models using rigorous evaluation procedures.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"7 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148893526","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comparative Field Evaluation of Active and Passive Transient‐Based Techniques for Pipeline Condition Assessment","authors":"Sirui Wang,Zhao Li,Pedro Lee,Tong‐Chuan Che,Yun‐jie Li,Ling Zhou","doi":"10.1029/2026wr043403","DOIUrl":"https://doi.org/10.1029/2026wr043403","url":null,"abstract":"Abstract Hydraulic transient‐based techniques are increasingly applied for non‐intrusive condition assessment of buried water pipelines. This study presents systematic field experiments comparing active and passive approaches within the asbestos cement water supply network at the University of Canterbury. In the active approach, controlled valve closures generated hydraulic transients analyzed via time‐domain reflectometry, while the passive approach utilized cross‐ and auto‐correlation of ambient pressure noise. Their performance was evaluated in terms of operational range, detection accuracy, and measurement duration. Results demonstrate that, for wave speed estimation, the passive approach enabled reliable time‐of‐flight estimation over a substantially greater sensor‐to‐sensor separation (projected maximum range of 980 m) than the active approach (459 m), while maintaining comparable accuracy (2%–3% error in time‐of‐flight). However, the active approach showed better robustness than the passive approach in identifying reflections from branches, valves, and joints, although the corresponding regression analysis showed only moderate performance. Measurement duration beyond 5 min showed only marginal benefit for passive wave speed estimation and negligible improvement for pipeline anomaly identification. Numerical simulations further clarified unexpected reflection features and provided insight into complex wave interactions. These findings highlight the complementary strengths of active and passive approaches and their potential for practical application in real‐world water pipeline monitoring.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"3 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148932877","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Multidimensional Critical Regulation Approach to Water Resources Optimization","authors":"Huanghao Yu,Binquan Li,Yong Zhao,Qingming Wang,Zeling Ren,Kuang Li,Zhongmin Liang","doi":"10.1029/2025wr043144","DOIUrl":"https://doi.org/10.1029/2025wr043144","url":null,"abstract":"Abstract Conventional water allocation models focus on reducing water shortage, maximizing economic benefits, and controlling environmental impacts, but rarely incorporate the coordinated evolution of multidimensional water systems. To address this gap, an integrated prediction‐simulation‐optimization framework was developed for 62 prefecture‐level cities in the Yellow River Basin, embedding multidimensional water‐system order directly into the optimization objective. Sectoral water use was estimated using the quota method, water availability was predicted with machine learning, and land‐use change was simulated using the PLUS model. Based on synergistic theory and information entropy, order degree was quantified across water resources, water environment, and water ecology subsystems, and the improved model was solved using NSGA‐II with entropy‐weighted TOPSIS. Water‐demand uncertainty was represented using sector‐specific truncated normal distributions and evaluated through Monte Carlo simulation with robust solution screening. Support vector machine achieved the best prediction performance, with 98% of cities meeting the accuracy threshold. Land‐use simulation yielded an average Kappa coefficient of 0.82; the most intense transitions during 2023–2030 occurred among grassland, cropland, and forestland. By 2030, both models project persistent water shortages in several cities. The improved model introduces total entropy‐increase violation as a fourth objective, thereby penalizing city‐level deterioration in water‐system order while maintaining balanced performance under conventional objectives. Robust solutions further reduce expected and worst‐case shortages under uncertainty, although basin‐scale deficits remain. This study moves water‐system order from post‐evaluation to an explicit optimization objective, providing decision support for coordinated water allocation under deterministic and uncertain conditions.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"66 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148932606","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"How can the Lateral Profile of Velocity Be Predicted on a Sloping Riverbank With Vegetation?","authors":"Yuqi Shan,Fujian Li,Sichen Sun,Ziqin Gui,Chao Liu","doi":"10.1029/2025wr042387","DOIUrl":"https://doi.org/10.1029/2025wr042387","url":null,"abstract":"Abstract Riverbanks have laterally sloping banks and are favorable environments for aquatic vegetation. A vegetated sloping bank alters the lateral profile of streamwise velocity across the entire channel. However, the impacts of flow, vegetation and side sloping banks on the velocity profile are unclear, and predicting the velocity profile on a vegetated sloping bank is challenging. In this study, rigid cylinders were used to mimic bank vegetation, and laboratory experiments were performed to clarify the mechanisms through which the lateral profile of depth‐averaged velocity changes in a trapezoidal channel with a vegetated sloping bank. The side slope and vegetation are the two main factors controlling the velocities in the bare bed and vegetated regions, varying the lateral profiles of depth‐averaged velocity across the entire channel. By using the flow continuity equation and momentum equation, a depth‐averaged dimensionless equation was developed for a vegetated trapezoidal channel. The new governing equation could not be solved analytically in the vegetated sloping bank region, so a difference solution method was used. The velocity prediction results were consistent with the measurements. If a new channel has a geometry condition (the ratio of channel width to flow depth and the bank slope) and vegetation density similar with a calibrated channel, the new method combining the difference solution method could be applied to predict the velocity profile across the new channel. Overall, this study provides an innovative approach to solving computational models that cannot be solved analytically in natural river environments, which provides technical support for river restoration.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"63 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148932876","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"From Lumped to Spatially Distributed Hydrologic Modeling: A Data‐Driven Framework Evaluated Across North American Catchments","authors":"Jinyang Li,Kuolin Hsu,Soroosh Sorooshian,Dan Lu","doi":"10.1029/2025wr043068","DOIUrl":"https://doi.org/10.1029/2025wr043068","url":null,"abstract":"Abstract Data‐driven rainfall–runoff models have advanced rapidly, yet the majority of large‐scale applications still rely on lumped inputs that smooth out spatial variability in precipitation, temperature, and landscape properties. This simplification can introduce substantial biases in flood peaks, hydrograph timing, and water balance estimates. Here, we develop a deep learning framework that ingests spatially distributed (gridded) meteorological forcings and is evaluated across 921 catchments in North America using the CAMELS‐SPAT data set. We compare two strategies for fusing spatial and temporal information, namely a spatial‐to‐temporal (S2T) architecture and a temporal‐to‐spatial (T2S) architecture, and assess the added value of gridded catchment attributes and finer spatial resolutions. The T2S approach consistently achieves the best performance, increasing median Nash–Sutcliffe Efficiency (NSE) from 0.60 (lumped baseline) to 0.65 and Kling–Gupta Efficiency (KGE) from 0.63 to 0.69, with NSE improvements observed in 75.4% of catchments. Performance gains are consistent across catchment sizes and aridity regimes, with enhanced skill in low‐flow, high‐flow, and peak timing representations. Integrated Gradients (IG) analysis shows that the model learns physically coherent spatial behavior, highlighting grid cells near the outlet and identifying dominant meteorological drivers of peak runoff events. These results demonstrate the practical value of spatially distributed forcings for data‐driven hydrologic modeling and outline a pathway toward interpretable, large‐scale streamflow prediction.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"1 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148932879","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The Role of Assimilating Deep Learning‐Enhanced AMSR‐E/2 Observations in European Soil Moisture Reanalysis","authors":"Visakh Sivaprasad,Johannes Keller,Yorck Ewerdwalbesloh,Anne Springer,Harrie‐Jan Hendricks Franssen,Harry Vereecken,Carsten Montzka","doi":"10.1029/2026wr043561","DOIUrl":"https://doi.org/10.1029/2026wr043561","url":null,"abstract":"Abstract Soil moisture is an essential climate variable that controls the exchange of water and energy between land and atmosphere, directly impacting land surface feedbacks. Satellite‐derived soil moisture provides valuable large‐scale information, but long‐term applications are limited by sensor lifespans, retrieval noise, and systematic biases. Earth system reanalyses rely on data assimilation to merge these observations with model dynamics. The mentioned limitations of the observations can degrade assimilation performance by propagating artifacts into model states. Here, we present a European daily soil moisture reanalysis for 2003–2023 that explicitly addresses these limitations by pre‐processing satellite observations prior to assimilation. We assimilate a newly developed Deep Learning (DL)‐enhanced AMSR‐E/2 soil moisture data set that provides a seamless daily record from 2003 to 2023 by reducing retrieval artifacts while preserving spatiotemporal consistency. Residual systematic biases are mitigated through climatological bias correction, and observation uncertainty is characterized using spatially distributed uncertainty estimates derived from Triple Collocation Analysis. The processed observations are subsequently assimilated into the encore Community Land Model (eCLM). This study evaluates whether DL‐enhanced observations improve reanalysis skill when assimilated into a continental‐scale ensemble land surface DA system. Evaluation against independent in situ measurements shows that assimilation of the DL‐enhanced AMSR‐E/2 reduces PBIAS from 28.68% to 3.28% and RMSE from 0.100 to 0.082 relative to open‐loop simulation, and outperforms assimilation of the original AMSR‐E/2 (PBIAS 15.13%, RMSE 0.094 ). Representing observation uncertainty as a spatially distributed field rather than a uniform value further improved temporal correlation at 66% of validation stations in a dedicated 3‐year experiment (2016–2018). Improvements are consistent across most in situ networks, with the largest gains observed in semi‐arid and transitional climate zones. At the same time, seasonal benefits are most pronounced during spring, autumn, and summer, where model dry‐down and vegetation dynamics are better represented. These improvements result in better preservation of hydrological signals and more physically consistent soil moisture dynamics. Our results demonstrate that observation pre‐processing is essential for effective soil moisture data assimilation, providing a reliable pathway for improved Earth system reanalyses.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"18 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148893696","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Probabilistic Pathway Framework for River–Sea Connectivity in Distributary Networks: A Case Study in the Guangdong–Hong Kong–Macao Greater Bay Area","authors":"Y. Z. Cai,M. L. Feng,T. Xie,Z. H. Ning,Q. Liu,F. Gao,J. Li,X. Li,B. S. Cui","doi":"10.1029/2025wr042155","DOIUrl":"https://doi.org/10.1029/2025wr042155","url":null,"abstract":"Abstract Hydrological connectivity regulates river–sea exchange by controlling freshwater delivery, salinity intrusion, sediment transport, and biogeochemical fluxes that sustain deltaic ecosystems. In distributary river systems, connectivity is mediated by multiple alternative downstream pathways, yet most assessments rely on network‐averaged indices and overlook cumulative regulation along interacting routes. Here, we develop a pathway‐inclusive, probabilistic framework to quantify river‐sea hydrological connectivity in distributary networks by retaining all feasible pathways and cumulative low‐head barrier effects. This formulation distinguishes structural redundancy from functional connectivity by accounting for the unequal contributions of different pathways to river–sea exchange. We apply the framework to the Guangdong–Hong Kong–Macao Greater Bay Area using multi‐decadal river network and barrier data sets from the 1980s to the 2010s. Total river network length increased from 7.9 × 10 6  m to 8.8 × 10 6  m, while sluice number increased more than sixfold and density twentyfold. Despite network expansion, functional river–sea connectivity declined by approximately 40% across maximum, average, and minimum connectivity metrics. The sharpest decline in average connectivity occurred between the 1990s and 2000s, indicating a system‐wide transition toward reduced pathway effectiveness. Connectivity losses were most pronounced along high‐capacity pathways that historically supported disproportionate exchange, whereas increased pathway number and small‐channel construction failed to offset network‐scale degradation. Spatially, connectivity declined most strongly upstream, despite barrier clustering near estuarine outlets, revealing non‐local, system‐wide regulatory effects. These results demonstrate that explicitly evaluating all feasible downstream pathways is necessary to diagnose connectivity loss, identify critical routes for river–sea exchange, and avoid misleading inferences based on network‐averaged indices.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"76 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148893533","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}