Water Resources Research最新文献

{"title":"Regional Variations in Drivers of Extreme Reference Evapotranspiration Across the Contiguous United States","authors":"Bingjie Zhao, Christopher Horvat, Christopher Pearson, Deep Shah, Huilin Gao","doi":"10.1029/2025wr040177","DOIUrl":"https://doi.org/10.1029/2025wr040177","url":null,"abstract":"Extremely high reference evapotranspiration indicates an abnormal atmospheric water demand, with significant implications for regional water resource management. Despite its importance, our understanding of extreme reference evapotranspiration remains limited, particularly regarding the contributions of its underlying drivers. This study focuses on all 339 hydrological unit code 6 basins across the Contiguous United States (CONUS) and, using two different methods for identifying extreme events, analyzes the meteorological contributions in various regions. We evaluate how factors such as meteorological data sets, meteorological variables, temporal and spatial scales, and the severity of extreme events introduce uncertainty into the analysis. The results reveal distinct regional patterns in the dominant drivers of extreme daily reference evapotranspiration across the CONUS. Air temperature is the dominant driver in the northern US, while solar radiation primarily drives extreme events in the southeastern US. Wind speed is the main driver in the southwestern US, with its influence increasing at higher severity levels or with finer temporal resolution. The air temperature and humidity jointly dominate the central US. All contributions from meteorological forcings vary with selection of temporal scales and severity levels, and utilizing multiple data sets enhances the robustness of extreme event identification.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"93 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2025-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145195075","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":"Denoising Autoencoder for Reconstructing Sensor Observation Data and Predicting Evapotranspiration: Noisy and Missing Values Repair and Uncertainty Quantification","authors":"Timothy K. Johnsen, Xiangyu Bi, Chunwei Chou, Charuleka Varadharajan, Yuxin Wu, Jonathan Skone, Lavanya Ramakrishnan","doi":"10.1029/2024wr039831","DOIUrl":"https://doi.org/10.1029/2024wr039831","url":null,"abstract":"Machine learning (ML) methods applied in scientific research often deal with interrelated features in high‐dimensional data. Reducing data noise and redundancy is needed to increase prediction accuracy and efficiency especially when dealing with data from field sensors. We explored an unsupervised learning method, the denoising autoencoder (DAE), to extract the underlying data structure from noisy raw data in the context of predicting hydrologic quantities from multiple field sensors. These sensors have intrinsic instrumental noise and occasional malfunctions that cause missing values. Our DAE neural network reconstructed meteorological sensor data containing noise and missing values to predict evapotranspiration in a mountainous watershed. The DAE reconstructed the sensor variables with a mean coefficient of determination value of 0.77 across 15 dimensions representing individual sensors. It reduced variance and bias uncertainties compared to a classical autoencoder model. The reconstruction quality varied across dimensions depending on their cross‐correlation and alignment with the underlying data structure. Uncertainties arising from the model structure were overall higher than those resulting from data corruption. We attached the DAE structure to a downstream ET‐prediction neural network in three formats and achieved reasonably accurate ET predictions . The use of the DAE notably reduced variance uncertainty in ET prediction. However, excessive variance reduction may be accompanied by an increase in bias due to the intrinsic bias‐variance tradeoff. Our method of evaluating and reducing uncertainties in aggregated data from different sources can be used to improve predictive models, process understanding, and uncertainty quantification for better water resource management.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"28 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2025-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145195057","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":"Flow Resistance Decomposition in the Presence of Leafy Flexible Vegetation and Sand Dunes","authors":"G. Artini, S. Francalanci, L. Solari, J. Aberle","doi":"10.1029/2025wr039942","DOIUrl":"https://doi.org/10.1029/2025wr039942","url":null,"abstract":"In open‐channel flows, hydraulic resistance is influenced by various factors, including sediment, channel geometry, and vegetation. Understanding how total resistance partitions into skin friction and form drag is fundamental for improving sediment transport predictions and advancing river morphodynamics knowledge. This study investigates the composition of total bed shear stress in environments featuring leafy flexible vegetation and sand dunes. Laboratory experiments were conducted under both mobile‐bed and fixed‐bed conditions using artificial plants with removable leafy branches. For mobile‐bed experiments, bed shear stress components were predicted using literature models for skin friction and form drag associated with bedforms and vegetation, whereas for fixed‐bed experiments they were derived from drag measurements. Results showed that the linear superposition principle fails when leafy vegetation is present. In such cases, total bed shear stress, estimated using the depth‐slope product, deviated, on average, by 52% under mobile‐bed conditions and 35% under fixed‐bed conditions from the sum of the individual stress components. In contrast, deviations averaged 15% in leafless setups. Under fixed‐bed conditions, total bed shear stress, inferred from drag measurements, exceeded model predictions by a factor of 2.2–3.3. Although the same dune model was used in all fixed‐bed setups, resistance coefficient associated with dune‐induced form drag increased exponentially with vegetation roughness density. This indicates existing models may underestimate dune‐related drag when leafy vegetation is present. Results highlight the role of foliage configuration in controlling total bed shear stress through non‐linear interactions with dune‐related form drag and indicate the need for predictive models accounting for such coupled effects.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"100 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145188962","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":"CCTV‐Hyperspectral Imaging for Suspended Sediment Transport (HISST): Proof‐of‐Concept for a Continuous Day‐and‐Night Monitoring Approach","authors":"Siyoon Kwon, Hyoseob Noh, Il Won Seo, Yun Ho Lee","doi":"10.1029/2025wr040402","DOIUrl":"https://doi.org/10.1029/2025wr040402","url":null,"abstract":"Effective sediment monitoring is crucial for managing dynamic river environments where suspended sediment transport varies over time. However, manual sampling and turbidity sensor‐based methods provide limited spatial coverage and can be labor‐intensive. Remote sensing offers non‐contact spatial measurements but generally has low temporal resolution. To overcome these challenges, we propose closed‐circuit television‐hyperspectral imaging for suspended sediment transport (CCTV‐HISST). This framework consists of a hyperspectral CCTV system integrated with a machine learning framework and enables continuous, high‐frequency monitoring of suspended sediment concentration (SSC) during the daytime, at sunset, and overnight. Combining hyperspectral imaging with low‐light adaptability, the system can detect subtle spectral variations in sediments under natural and artificial lighting. We conducted 15 experiments using three sediment types (high‐visibility silt, low‐visibility sand, and their mixture) under controlled shallow‐water conditions in an outdoor flume. Experiments were categorized by light source: sunlight for daytime, combined sunlight and halogen lighting at sunset, and halogen lighting at night. This proof‐of‐concept study suggests that the proposed machine learning framework, light classification and adaptive regression, achieved 99% accuracy in light classification and strong agreement with field SSC measurements, even in untrained cases. Validation using field spectrometry and laser diffraction sensor data further confirmed the reliability of the proposed system. This study highlights the potential of CCTV‐HISST as a scalable, noncontact alternative for real‐time monitoring by adaptively detecting suspended sediments and quantifying their concentration across a range of light conditions. Future studies can extend its applicability to natural rivers by addressing limitations related to water depth and SSC variability.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"54 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145188964","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":"Improving Streamflow Predictions in the Arid Southwestern United States Through Understanding of Baseflow Generation Mechanisms","authors":"Mohammad A. Farmani, Ahmad Tavakoly, Ali Behrangi, Yuan Qiu, Aniket Gupta, Muhammad Jawad, Hossein Yousefi Sohi, Xueyan Zhang, Matthew Geheran, Guo‐Yue Niu","doi":"10.1029/2024wr039479","DOIUrl":"https://doi.org/10.1029/2024wr039479","url":null,"abstract":"Understanding the factors controlling baseflow (groundwater discharge) is critical for improving streamflow predictions in the arid southwestern United States. We used an enhanced version of the Noah‐MP land surface model with advanced hydrological process options and the Routing Application for Parallel computation of Discharge (RAPID) to examine the impacts of process representation, soil hydraulic parameters, and precipitation data sets on baseflow production and streamflow skill. Model experiments combined multiple configurations of hydrological processes, soil parameters, and three gridded precipitation products: NLDAS‐2, Integrated Multi‐satellite Retrievals for GPM Final, and NOAA AORC. RAPID was used to route Noah‐MP‐simulated runoff and generate daily streamflow at 390 U.S. Geological Survey (USGS) gauges. The modeled baseflow index (BFI) was compared with USGS‐derived BFI. Results show that (a) soil water retention curve model plays a dominant role, with the Van‐Genuchten hydraulic scheme reducing the overestimated BFI produced by the Brooks‐Corey, (b) hydraulic parameters (Van‐Genuchten parameters and hydraulic conductivity) strongly affect streamflow prediction, a machine learning‐based Van‐Genuchten parameters captures the USGS BFI, showing a better performance than the optimized National Water Model (NWM) by a median Kling‐Gupta Efficiency of 21%, and (c) incorporating a ponding depth threshold into the land surface models that increases infiltration is preferred. Overall, models with more physically realistic hydrologic representations show a better performance in modeling BFI and thus a better skill in streamflow predictions than the optimized NWM in the dry southwestern river basins. These findings can guide future studies in selecting reliable schemes and data sets (before calibration) to achieve better streamflow predictions as well as water resource projections.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"104 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145189029","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":"Combined Measurement of Soil Permittivity and Electrical Conductivity Using UAV‐Based Ground Penetrating Radar","authors":"Qinbo Cheng, Yu Cheng, Zhijin Ma, Andrew Binley, Jintao Liu, Zhicai Zhang, Feng Huang, Xi Chen","doi":"10.1029/2024wr039519","DOIUrl":"https://doi.org/10.1029/2024wr039519","url":null,"abstract":"Measurements of soil water content and salinity are important for a wide range of topics, in particular those concerned with soil and plant health, and specific aspects of agricultural management. However, most traditional methods are unsuitable for simultaneously mapping the field scale variability of soil electrical properties. In this study, we propose a method that uses an unmanned aerial vehicle (UAV) to support ground penetrating radar (GPR) antennae with different frequencies, allowing spatial scanning of surface reflection coefficients, which is then used to estimate the soil relative permittivity (<jats:italic>ε</jats:italic><jats:sub><jats:italic>r</jats:italic></jats:sub>) and electrical conductivity (<jats:italic>σ</jats:italic>). These parameters are then used to estimate soil water content and salinity using empirical transfer functions. Unlike other published approaches, the proposed method is relatively simple and does not rely on full‐waveform inversion. Field tests in the riparian zone of the Yangtze River and salinized land close to the Yellow Sea are used to demonstrate the effectiveness of the method. The surveys illustrate that the UAV‐GPR give results comparable to those measured in situ with a soil electrical property meter. These findings are supported by accuracy analysis using Monte Carlo simulation which reveal that the measurement error of <jats:italic>ε</jats:italic><jats:sub><jats:italic>r</jats:italic></jats:sub> increases with <jats:italic>σ</jats:italic>, and the relative errors in <jats:italic>σ</jats:italic> measurements are generally less than those of <jats:italic>ε</jats:italic><jats:sub><jats:italic>r</jats:italic></jats:sub> except in areas of high <jats:italic>ε</jats:italic><jats:sub><jats:italic>r</jats:italic></jats:sub> and low <jats:italic>σ</jats:italic>. The study provides an approach for mapping soil electrical properties using UAV technology, thus opening up the possibility of remote sensing of spatial variability of these important properties at high spatial resolution.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"6 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145188963","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":"Coupled Hydrologic and Biogeochemical Responses of Nitrate Export in a Tile‐Drained Agricultural Watershed Revealed by SAS Functions and Nitrate Isotopes","authors":"Minghui Sha, Zhongjie Yu, Paolo Benettin, Lowell E. Gentry, Corey A. Mitchell","doi":"10.1029/2024wr039718","DOIUrl":"https://doi.org/10.1029/2024wr039718","url":null,"abstract":"The combination of high nitrogen (N) inputs on tile‐drained agricultural watersheds contributes to excessive nitrate (NO<jats:sub>3</jats:sub><jats:sup>−</jats:sup>) loss to surface‐ and groundwater systems. This study combined water age modeling based on StorAge Selection functions and NO<jats:sub>3</jats:sub><jats:sup>−</jats:sup> isotopic analysis to examine the underlying mechanisms driving NO<jats:sub>3</jats:sub><jats:sup>−</jats:sup> export in an intensively tile‐drained mesoscale watershed typical of the U.S. Upper Midwest. The water age modeling revealed a pronounced inverse storage effect and strong young water preference under high‐flow conditions, emphasizing evolving water mixing behavior driven by groundwater fluctuation and tile drain activation. Integrating NO<jats:sub>3</jats:sub><jats:sup>−</jats:sup> concentration‐isotope‐discharge relationships with water age dynamics disentangled the interactions between flow path variations and subsurface N cycling in shaping seasonally variable NO<jats:sub>3</jats:sub><jats:sup>−</jats:sup> export regimes at the watershed scale. Based on these results, a simple transit time‐based and isotope‐aided NO<jats:sub>3</jats:sub><jats:sup>−</jats:sup> transport model was developed to estimate the timescales of watershed‐scale NO<jats:sub>3</jats:sub><jats:sup>−</jats:sup> reactive transport. Model results demonstrated variable NO<jats:sub>3</jats:sub><jats:sup>−</jats:sup> source availability and a wetness dependence for denitrification, indicating that interannual NO<jats:sub>3</jats:sub><jats:sup>−</jats:sup> chemostasis is driven by coupled and proportional responses of soil NO<jats:sub>3</jats:sub><jats:sup>−</jats:sup> production, denitrification, and flow path activation to varying antecedent wetness conditions. These findings suggest that intensively tile‐drained Midwestern agricultural watersheds function as both N transporters and transformers and may respond to large‐scale mitigation efforts within a relatively short timeframe. Collectively, the results of this study demonstrate the potential of integrated water age modeling and NO<jats:sub>3</jats:sub><jats:sup>−</jats:sup> isotopic analysis to advance the understanding of macroscale principles governing coupled watershed hydrologic and N biogeochemical functions.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"32 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145189030","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":"Effective Thermal Retardation in Aquifers of Heterogeneous Hydraulic Conductivity","authors":"Hannah Gebhardt, Alraune Zech, Gabriel C. Rau, Peter Bayer","doi":"10.1029/2025wr040153","DOIUrl":"https://doi.org/10.1029/2025wr040153","url":null,"abstract":"Thermal retardation and dispersion are important processes affecting advective heat transport in sedimentary aquifers, yet little is known how they are influenced by heterogeneity of hydraulic conductivity. We investigate the effect of macro‐scale heterogeneity on transient heat transport in a three‐dimensional domain through direct numerical Monte‐Carlo simulations. The model describes the evolution of a heat plume in a heterogeneous aquifer generated by a borehole heat exchanger. We characterize the transport by calculating the dispersion coefficient and effective thermal retardation factor as ensemble average of the heterogeneous realizations. In addition to different degrees of heterogeneity, we examine the influence of the thermal Péclet number on the effective thermal retardation factor. Simulations reveal that for homogeneous hydraulic conductivity, the effective thermal retardation factor equals the predicted, apparent thermal retardation factor. However, in heterogeneous cases, the effective thermal retardation factor is substantially lower than the apparent value at early times, with this effect becoming more pronounced as the Péclet number increases. We attribute the deviation of the effective thermal retardation factor from the apparent value to preferential flow through zones with higher hydraulic conductivity and delayed local heat diffusion into zones with lower hydraulic conductivity. Assuming that the effective thermal retardation factor differs from the apparent value in the presence of local thermal non‐equilibrium (LTNE) effects, we call the observed effect “field‐scale LTNE.” Finally, we derive a formula estimating effective thermal retardation as a function of log‐conductivity variance and the Péclet number. Our results can improve heat tracer techniques in hydraulically heterogeneous environments.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"31 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145189518","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":"Experimental Warming and Short‐Term Nutrient Effects on Nearshore Periphyton Metabolism in a Large, Oligotrophic Lake","authors":"Nicholas T. Framsted, Adrianne P. Smits, Steven Sadro","doi":"10.1029/2025wr039891","DOIUrl":"https://doi.org/10.1029/2025wr039891","url":null,"abstract":"Periphyton blooms may be increasing in oligotrophic lakes due to warming water temperatures and increased nutrient loads associated with climate change. Such blooms decrease both water quality and the aesthetic value of nearshore areas, but identifying the mechanisms driving periphyton blooms in situ is complex. We conducted laboratory experiments using periphyton‐covered rocks collected from the nearshore of oligotrophic Lake Tahoe, CA, to examine (a) baseline seasonal variability in periphyton biomass and metabolism, (b) effects of warming, nutrients, and their interaction on periphyton metabolism, and (c) seasonal variability in these effects on periphyton metabolism. We quantified rates of gross primary production (GPP), ecosystem respiration (ER), and net ecosystem production (NEP) under 2 nutrient treatments (ambient and enriched) and 4 warming treatments (ambient, +3°C, +6°C, and +9°C above ambient). Overall, warming stimulated GPP, NEP, and ER (Q<jats:sub>10</jats:sub> temperature coefficients of 1.6, 1.4, and 2 respectively), with stronger effects during colder months. Warming also stimulated metabolic rates in the absence of nutrient additions. Short‐term nutrient effects were more variable across seasons and alternated between depressing or stimulating metabolic rates (ranged from a 9%–11% decrease in rates to a 13%–27% increase across seasons). The relative importance of warming and nutrient effects were seasonally dependent as nutrients stimulated metabolic rates more than warming in October, and warming more so than nutrients in February and November. These results indicate that climate‐driven alterations to temperature and nutrient regimes will have important and seasonally explicit consequences for the ecosystem energetics and periphyton community structure of oligotrophic lakes.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"23 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145189028","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":"Soil Physics‐Informed Neural Networks to Estimate Bimodal Soil Hydraulic Properties","authors":"Jieliang Zhou, Yunquan Wang, Pengfei Qi, Rui Ma, Harry Vereecken, Budiman Minasny, Yonggen Zhang","doi":"10.1029/2024wr039337","DOIUrl":"https://doi.org/10.1029/2024wr039337","url":null,"abstract":"Pedotransfer functions (PTFs) are widely used to estimate soil hydraulic properties (SHPs) from easily measurable soil properties. However, most existing PTFs are based on unimodal hydraulic models, which fail to capture the bimodal behavior of hydraulic properties caused by soil structure. In this study, we developed new PTFs using two bimodal hydraulic models and introduced a soil physics–informed neural network to embed these models into the training process. The results showed that the new PTFs effectively represented bimodality in hydraulic conductivity curves, achieving a root mean square error of 0.531(<jats:italic>K</jats:italic> in cm/d) on the test set, compared with 0.626 for unimodal models. They also improved predictions of soil water retention curves but struggled with bimodal cases for some samples, likely due to the limited number of bimodal retention curves in the training data set. Evaluation on an independent data set showed that the error for hydraulic conductivity predicted by the new functions was about one‐third that of conventional approaches. In addition, the proposed soil physics–informed neural network, which directly optimizes SHPs, outperformed the conventional approaches that optimize fitted model parameters. We also found that whether water retention and hydraulic conductivity are optimized separately or simultaneously has a large effect on performance. Nonetheless, the lack of explicit soil structure information in the input data, along with limited measurements near saturation, continues to constrain accuracy. This emphasizes the need to develop a more comprehensive soil hydraulic database.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"2 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145189103","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}