Zili Wang, Chaoyue Li, Ruilong Wei, Binlan Zhang, Peng Cui
Hydrological signatures (HS) have proven to be highly effective in calibrating physically-based hydrological models, enhancing their process consistency. However, their integration into parameter optimization for deep learning (DL)-based hydrological models has been limited. To address this gap, we propose a novel HS-informed framework that dynamically integrates HS into DL parameterization through a multi-task learning approach. This study evaluates the impact of HS integration on model performance using a large-scale, global hydrological data set. The HS-informed model achieved a significant performance improvement, with a median Nash-Sutcliffe Efficiency (NSE) of 0.739, compared to 0.666 for the baseline model across the test set. Notably, the most pronounced improvements in NSE were observed in hydrologically complex basins, including baseflow-dominated (+0.135), drought-prone (+0.148), and flood-prone basins (+0.159). Sensitivity analysis further revealed that the HS-informed model could leverage extended historical input data (over 120 days) to sustain robust performance (median NSE of 0.715) over a 30-day forecast period. Shapley Additive Explanations analysis highlighted two key mechanisms underlying these improvements: the enhanced recognition of long-term hydrological patterns through improved memory and a better representation of catchment heterogeneity by emphasizing non-climatic attributes. These findings demonstrate that integrating HS offers a superior approach to traditional point-error-based calibration in AI-driven hydrological modeling.
{"title":"A Novel Hydrological Signature-Informed Framework for Enhancing Streamflow Prediction Using Multi-Task Learning","authors":"Zili Wang, Chaoyue Li, Ruilong Wei, Binlan Zhang, Peng Cui","doi":"10.1029/2025wr041485","DOIUrl":"https://doi.org/10.1029/2025wr041485","url":null,"abstract":"Hydrological signatures (HS) have proven to be highly effective in calibrating physically-based hydrological models, enhancing their process consistency. However, their integration into parameter optimization for deep learning (DL)-based hydrological models has been limited. To address this gap, we propose a novel HS-informed framework that dynamically integrates HS into DL parameterization through a multi-task learning approach. This study evaluates the impact of HS integration on model performance using a large-scale, global hydrological data set. The HS-informed model achieved a significant performance improvement, with a median Nash-Sutcliffe Efficiency (NSE) of 0.739, compared to 0.666 for the baseline model across the test set. Notably, the most pronounced improvements in NSE were observed in hydrologically complex basins, including baseflow-dominated (+0.135), drought-prone (+0.148), and flood-prone basins (+0.159). Sensitivity analysis further revealed that the HS-informed model could leverage extended historical input data (over 120 days) to sustain robust performance (median NSE of 0.715) over a 30-day forecast period. Shapley Additive Explanations analysis highlighted two key mechanisms underlying these improvements: the enhanced recognition of long-term hydrological patterns through improved memory and a better representation of catchment heterogeneity by emphasizing non-climatic attributes. These findings demonstrate that integrating HS offers a superior approach to traditional point-error-based calibration in AI-driven hydrological modeling.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"15 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-01-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145971983","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}
The altitude effect (AE) of stable isotopes in meteoric water (δ18O and δ2H), that is, the depletion of water isotopes with increasing altitude, is an important theoretical assumption of isotope-based paleoaltimetry. However, this assumption has recently been challenged, as many in situ observations fail to consistently demonstrate the expected negative correlation between altitude and isotope values. Here we used 1,255 records of surface water isotopes to investigate AE and inverse altitude effect (IAE) and their mechanisms in arid Central Asia. The results show that isotope altitude gradients across Central Asia are weaker than the global average. Comparisons of the gradients for both the mountain-basin system and mountain system reveal that the windward and leeward slopes of the westerlies consistently exhibit opposite gradients: AE on the windward side and IAE on the leeward. The observed IAE on the leeward slope across all basins is influenced by topography and local circulation. The orientation of mountain ranges perpendicular to large-scale westerly circulation blocks eastward transport of westerly moisture, and the resulting longer moisture pathways weaken AE. Stronger local circulation and sub-cloud evaporation processes enrich water isotopes in the leeward mountain regions, diminishing AE and even leading to the emergence of IAE. Our results highlight the impact of local circulation on water isotopes during different uplift phases when using stable hydrogen and oxygen isotopes to reconstruct paleoelevation.
{"title":"Weakened Isotope Altitude Gradient in the Central Asian Water Tower: Role of Topography and Local Circulation","authors":"Yudong Shi, Shengjie Wang, Xiaokang Liu, Kei Yoshimura, Hayoung Bong, Chenggang Zhu, Yanjun Che, Huawu Wu, Mingjun Zhang","doi":"10.1029/2025wr040283","DOIUrl":"https://doi.org/10.1029/2025wr040283","url":null,"abstract":"The altitude effect (AE) of stable isotopes in meteoric water (<i>δ</i><sup>18</sup>O and <i>δ</i><sup>2</sup>H), that is, the depletion of water isotopes with increasing altitude, is an important theoretical assumption of isotope-based paleoaltimetry. However, this assumption has recently been challenged, as many <i>in situ</i> observations fail to consistently demonstrate the expected negative correlation between altitude and isotope values. Here we used 1,255 records of surface water isotopes to investigate AE and inverse altitude effect (IAE) and their mechanisms in arid Central Asia. The results show that isotope altitude gradients across Central Asia are weaker than the global average. Comparisons of the gradients for both the mountain-basin system and mountain system reveal that the windward and leeward slopes of the westerlies consistently exhibit opposite gradients: AE on the windward side and IAE on the leeward. The observed IAE on the leeward slope across all basins is influenced by topography and local circulation. The orientation of mountain ranges perpendicular to large-scale westerly circulation blocks eastward transport of westerly moisture, and the resulting longer moisture pathways weaken AE. Stronger local circulation and sub-cloud evaporation processes enrich water isotopes in the leeward mountain regions, diminishing AE and even leading to the emergence of IAE. Our results highlight the impact of local circulation on water isotopes during different uplift phases when using stable hydrogen and oxygen isotopes to reconstruct paleoelevation.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"41 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-01-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145972440","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}
Groundwater flow and solute transport models, governed by partial differential equations (PDEs), are computationally intensive, particularly in large-scale. Traditional numerical models are prohibitively expensive, and existing surrogate models often fail under out-of-distribution (OOD) conditions, such as unseen initial conditions, boundary configurations or altered source terms. To address these challenges, we propose a novel framework based on Operator Inference (OpInf), a physics-informed surrogate modeling approach. OpInf preserves the structure of governing equations, ensuring physical consistency and interpretability, while significantly improving computational efficiency and generalization capabilities. By leveraging Proper Orthogonal Decomposition (POD) for dimensionality reduction and inferring reduced operators directly from simulation data, OpInf enables robust prediction of system behavior. We evaluate the proposed method through two case studies: the two-dimensional and three-dimensional solute transport problem under different point-source concentration fluctuation release conditions with heterogeneous hydraulic conductivity. The inversion framework is further appraised by integrating OpInf with Wasserstein Generative Adversarial Networks with Gradient Penalty (WGAN-GP) parameterization methods and the Ensemble Smoother (ES) data assimilation. Results demonstrate that OpInf relatively outperforms other surrogate models, particularly under OOD conditions and the inversion efficiency can be increased by over 99%. We establish OpInf as a transformative tool for dynamic surrogate groundwater modeling, offering robust generalization, reduced computational costs, and strong potential for real-world applications.
{"title":"Operator Inference for Physical and Generalized Surrogate Groundwater Modeling","authors":"Yongda Liu, Xi Chen, Zitao Wang, Jianzhi Dong","doi":"10.1029/2025wr039961","DOIUrl":"https://doi.org/10.1029/2025wr039961","url":null,"abstract":"Groundwater flow and solute transport models, governed by partial differential equations (PDEs), are computationally intensive, particularly in large-scale. Traditional numerical models are prohibitively expensive, and existing surrogate models often fail under out-of-distribution (OOD) conditions, such as unseen initial conditions, boundary configurations or altered source terms. To address these challenges, we propose a novel framework based on Operator Inference (OpInf), a physics-informed surrogate modeling approach. OpInf preserves the structure of governing equations, ensuring physical consistency and interpretability, while significantly improving computational efficiency and generalization capabilities. By leveraging Proper Orthogonal Decomposition (POD) for dimensionality reduction and inferring reduced operators directly from simulation data, OpInf enables robust prediction of system behavior. We evaluate the proposed method through two case studies: the two-dimensional and three-dimensional solute transport problem under different point-source concentration fluctuation release conditions with heterogeneous hydraulic conductivity. The inversion framework is further appraised by integrating OpInf with Wasserstein Generative Adversarial Networks with Gradient Penalty (WGAN-GP) parameterization methods and the Ensemble Smoother (ES) data assimilation. Results demonstrate that OpInf relatively outperforms other surrogate models, particularly under OOD conditions and the inversion efficiency can be increased by over 99%. We establish OpInf as a transformative tool for dynamic surrogate groundwater modeling, offering robust generalization, reduced computational costs, and strong potential for real-world applications.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"101 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-01-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146005714","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}
Intensifying flooding throughout the western United States threatens human infrastructure, human life, and ecological integrity. Flash floods are particularly dangerous because they rise quickly and often unexpectedly. Floodplains that are hydrologically connected to river channels can act as buffers to attenuate peak flows and slow flood movement, providing natural resilience against flood risks. Comparatively low-gradient reaches (1%–3% slopes) with wide floodplains (beads) have been identified as important for attenuating floods, yet the degree to which they can reduce peak flows is not well constrained. Additionally, the relative importance of attenuation mechanisms have not been discerned. We quantified flood attenuation provided by beads by utilizing two-dimensional hydrodynamic models simulating flash floods in three river beads located in the Colorado Rocky Mountains, United States. We quantified the (a) magnitude of attenuation, (b) total accessible floodplain volume, (c) volume of floodwater stored in floodplain depressions, (d) variability of flow path travel times, (e) floodplain heterogeneity, and (f) relative importance of these mechanisms in flood attenuation. We found unprecedently high discharge attenuation with an average 13.8% reduction in peak flow per kilometer reach length and continued attenuation up to the 100-year recurrence interval flood. For the studied sites the strongest correlations were between attenuation and storage in floodplain depressions. Flow path diversity metrics correlated best with attenuation for floods with a time-to-peak greater than 1 hr. Our findings also indicated that maintenance of high floodplain roughness and accessibility may be an effective strategy for bolstering attenuation of flash floods in mountain systems.
{"title":"The Relative Importance of Floodplain Storage and Flow Path Dispersion on Flood Attenuation in Mountain Streams","authors":"Nicholas Christensen, Ryan R. Morrison","doi":"10.1029/2024wr039628","DOIUrl":"https://doi.org/10.1029/2024wr039628","url":null,"abstract":"Intensifying flooding throughout the western United States threatens human infrastructure, human life, and ecological integrity. Flash floods are particularly dangerous because they rise quickly and often unexpectedly. Floodplains that are hydrologically connected to river channels can act as buffers to attenuate peak flows and slow flood movement, providing natural resilience against flood risks. Comparatively low-gradient reaches (1%–3% slopes) with wide floodplains (beads) have been identified as important for attenuating floods, yet the degree to which they can reduce peak flows is not well constrained. Additionally, the relative importance of attenuation mechanisms have not been discerned. We quantified flood attenuation provided by beads by utilizing two-dimensional hydrodynamic models simulating flash floods in three river beads located in the Colorado Rocky Mountains, United States. We quantified the (a) magnitude of attenuation, (b) total accessible floodplain volume, (c) volume of floodwater stored in floodplain depressions, (d) variability of flow path travel times, (e) floodplain heterogeneity, and (f) relative importance of these mechanisms in flood attenuation. We found unprecedently high discharge attenuation with an average 13.8% reduction in peak flow per kilometer reach length and continued attenuation up to the 100-year recurrence interval flood. For the studied sites the strongest correlations were between attenuation and storage in floodplain depressions. Flow path diversity metrics correlated best with attenuation for floods with a time-to-peak greater than 1 hr. Our findings also indicated that maintenance of high floodplain roughness and accessibility may be an effective strategy for bolstering attenuation of flash floods in mountain systems.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"267 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145961848","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}
M. Van Appledorn, N. R. De Jager, J. J. Rohweder, M. Windmuller-Campione, D. Griffin
The hydrologic regime of the upper Mississippi River (UMR) has become wetter, with greater discharges, longer-lasting high-flow conditions, and seasonal shifts in these patterns over the past several decades. How these changes are expressed spatially as floodplain inundation area, frequency, depth, duration, and timing is not well understood. It is also unclear to what degree spatial patterns of submergence are represented by examining discharge data alone. We assessed changes in floodplain inundation characteristics from 1940 to 2022 in navigation pools 3–10 of the UMR using a geospatial model to simulate daily inundation depths. Inundation characteristics shifted significantly across pools, but the direction and magnitude of change varied by pool and metric. Characteristics summarized at the pool scale correlated with streamgage-derived proxies but the strength of the relationship varied. Within pools, variability in inundation trends highlighted the importance of spatially explicit modeling. Our study demonstrates that changes in discharge over 83 years have manifested across the UMR floodplain in ways that may have consequences for ecological patterns and processes. By mapping hydrologically sensitive areas, we can anticipate which areas may be susceptible to additional shifts in river discharge in a climatically uncertain future.
{"title":"More Water, More of the Time: Spatial Changes in Flooding Over 83 Years in the Upper Mississippi River Floodplain and Relationships With Streamgage-Derived Proxies","authors":"M. Van Appledorn, N. R. De Jager, J. J. Rohweder, M. Windmuller-Campione, D. Griffin","doi":"10.1029/2025wr040614","DOIUrl":"https://doi.org/10.1029/2025wr040614","url":null,"abstract":"The hydrologic regime of the upper Mississippi River (UMR) has become wetter, with greater discharges, longer-lasting high-flow conditions, and seasonal shifts in these patterns over the past several decades. How these changes are expressed spatially as floodplain inundation area, frequency, depth, duration, and timing is not well understood. It is also unclear to what degree spatial patterns of submergence are represented by examining discharge data alone. We assessed changes in floodplain inundation characteristics from 1940 to 2022 in navigation pools 3–10 of the UMR using a geospatial model to simulate daily inundation depths. Inundation characteristics shifted significantly across pools, but the direction and magnitude of change varied by pool and metric. Characteristics summarized at the pool scale correlated with streamgage-derived proxies but the strength of the relationship varied. Within pools, variability in inundation trends highlighted the importance of spatially explicit modeling. Our study demonstrates that changes in discharge over 83 years have manifested across the UMR floodplain in ways that may have consequences for ecological patterns and processes. By mapping hydrologically sensitive areas, we can anticipate which areas may be susceptible to additional shifts in river discharge in a climatically uncertain future.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"57 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145962157","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}
Current models on solute transport often fail to reproduce discharge-dependent behavior of solute transport in stream reaches because they rely on the assumption of well-mixed conditions and fail to account for the complex coupling between in-stream and subsurface flow. StorAge Selection (SAS) functions describe outflow as a mixture of waters of different ages, providing a framework to overcome the well-mixed assumption in “traditional” transport models. In this study, we applied SAS functions to model solute transport from 13 slug tracer experiments conducted under varying hydrological conditions in a headwater stream reach. Using SAS function parameters (expressed in units of volume) together with measurements of groundwater (GW) levels and streambed microtopography, we partitioned the total water storage within the study reach into distinct components: advective storage, in-stream transient storage, tracer-derived hyporheic storage, and GW level-derived hyporheic storage. This partitioning assumes that transport processes and subsurface water flow in stream reaches are associated with different storage volumes. We found positive linear relationships between discharge and age-ranked, advective, and tracer-derived hyporheic storage. In-stream transient storage increased with discharge up to 17 L s−1, corresponding to the discharge threshold above which streambed sediments became completely submerged, and declined at higher flows. This pattern likely reflects the contribution of eddies at lower discharge levels and highlights the importance of in-stream transient storage for solute transport. Our results demonstrate that partitioning the total water storage in a reach–enabled only through applying SAS functions–is essential for understanding and modeling solute transport under varying hydrological conditions.
目前的溶质输运模型往往不能再现河流河段中溶质输运的流量依赖行为，因为它们依赖于充分混合条件的假设，而不能考虑流内流和地下流之间的复杂耦合。储存选择（SAS）函数将流出水描述为不同年龄的水的混合物，为克服“传统”运输模型中的混合假设提供了一个框架。在这项研究中，我们应用SAS函数来模拟在不同水文条件下的13个段塞流示踪剂实验中的溶质运移。利用SAS函数参数（以体积单位表示）以及地下水水位和河床微地形测量，我们将研究范围内的总储水量划分为不同的组成部分：平流储水量、流内瞬时储水量、示踪剂衍生的地下储水量和GW水平衍生的地下储水量。这种划分假设了运输过程和河流河段的地下水流与不同的储存量有关。我们发现放电与年龄分级、平流和示踪剂衍生的低循环储存呈正线性关系。当流量达到17 L s−1时，河道内瞬时库存量增加，超过该流量，河床沉积物完全被淹没；这种模式可能反映了低流量水平下涡流的贡献，并突出了流内瞬态储存对溶质运输的重要性。我们的研究结果表明，划分河段的总储水量（仅通过应用SAS函数实现）对于理解和模拟不同水文条件下的溶质运移至关重要。
{"title":"Partitioning Water Storage in Stream Reaches: Implications for Solute Transport Under Varying Hydrological Conditions","authors":"C. Glaser, E. Bonanno, G. Blöschl, J. Klaus","doi":"10.1029/2025wr040372","DOIUrl":"https://doi.org/10.1029/2025wr040372","url":null,"abstract":"Current models on solute transport often fail to reproduce discharge-dependent behavior of solute transport in stream reaches because they rely on the assumption of well-mixed conditions and fail to account for the complex coupling between in-stream and subsurface flow. StorAge Selection (SAS) functions describe outflow as a mixture of waters of different ages, providing a framework to overcome the well-mixed assumption in “traditional” transport models. In this study, we applied SAS functions to model solute transport from 13 slug tracer experiments conducted under varying hydrological conditions in a headwater stream reach. Using SAS function parameters (expressed in units of volume) together with measurements of groundwater (GW) levels and streambed microtopography, we partitioned the total water storage within the study reach into distinct components: advective storage, in-stream transient storage, tracer-derived hyporheic storage, and GW level-derived hyporheic storage. This partitioning assumes that transport processes and subsurface water flow in stream reaches are associated with different storage volumes. We found positive linear relationships between discharge and age-ranked, advective, and tracer-derived hyporheic storage. In-stream transient storage increased with discharge up to 17 L s<sup>−1</sup>, corresponding to the discharge threshold above which streambed sediments became completely submerged, and declined at higher flows. This pattern likely reflects the contribution of eddies at lower discharge levels and highlights the importance of in-stream transient storage for solute transport. Our results demonstrate that partitioning the total water storage in a reach–enabled only through applying SAS functions–is essential for understanding and modeling solute transport under varying hydrological conditions.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"15 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145968875","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}
Pengfei Qi, Yunquan Wang, Rui Ma, Jieliang Zhou, Harry Vereecken, Budiman Minasny, Ziyong Sun, Gaofeng Zhu, Kun Zhang
Pedotransfer functions (PTFs) are widely used to estimate soil hydraulic parameters based on easily accessible soil information, playing an important role in the parameterization of earth surface models. However, conventional PTFs, developed using measurements from small volume soil samples, often exhibit significant deviations from field observations and substantial variability when applied to field-scale hydrological models. Here, we introduce new Site-Specific Pedotransfer Functions (SPTFs) that combine deep learning with physics-based modeling of soil hydrological processes. SPTFs differ from conventional PTFs in two aspects: they utilize time-series data as input and they directly optimize simulated soil water content by the 1-D Richardson–Richards equation with observations, ensuring improved applicability to field conditions. We trained and tested the model using two years of soil moisture observations from 1,181 sites in the International Soil Moisture Network. Evaluation using field data demonstrates that SPTFs achieve a Nash-Sutcliffe Efficiency of 0.65 and root mean squared error of 0.072 cm3 cm−3 in simulating soil water content at the depth of 0.05 m on the test set (n = 179), which is close to the values predicted by the inverse modeling method, while maintaining the computational efficiency of PTFs. This study highlights the promise of SPTFs as a robust parameterization framework for localized field applications.
土壤传递函数（PTFs）被广泛应用于基于易于获取的土壤信息估计土壤水力参数，在地表模型的参数化中起着重要作用。然而，利用小体积土壤样品的测量方法开发的传统ptf，在应用于现场尺度水文模型时，往往与现场观测结果有很大的偏差，并且存在很大的变异性。在这里，我们引入了新的场地特定土壤传递函数（SPTFs），该函数将深度学习与基于物理的土壤水文过程建模相结合。sptf与传统PTFs的不同之处有两个方面：它们利用时间序列数据作为输入，并通过1-D Richardson-Richards方程与观测结果直接优化模拟土壤含水量，确保提高对现场条件的适用性。我们使用国际土壤湿度网络中1181个站点的两年土壤湿度观测数据对该模型进行了训练和测试。现场数据评价表明，在测试集（n = 179）上，sptf模拟0.05 m深度土壤含水量的Nash-Sutcliffe效率为0.65，均方根误差为0.072 cm3 cm - 3，在保持PTFs计算效率的前提下，与反建模方法预测值接近。这项研究强调了sptf作为本地化现场应用的鲁棒参数化框架的前景。
{"title":"Physics-Informed Neural Networks to Develop Site-Specific Pedotransfer Functions","authors":"Pengfei Qi, Yunquan Wang, Rui Ma, Jieliang Zhou, Harry Vereecken, Budiman Minasny, Ziyong Sun, Gaofeng Zhu, Kun Zhang","doi":"10.1029/2025wr041265","DOIUrl":"https://doi.org/10.1029/2025wr041265","url":null,"abstract":"Pedotransfer functions (PTFs) are widely used to estimate soil hydraulic parameters based on easily accessible soil information, playing an important role in the parameterization of earth surface models. However, conventional PTFs, developed using measurements from small volume soil samples, often exhibit significant deviations from field observations and substantial variability when applied to field-scale hydrological models. Here, we introduce new Site-Specific Pedotransfer Functions (SPTFs) that combine deep learning with physics-based modeling of soil hydrological processes. SPTFs differ from conventional PTFs in two aspects: they utilize time-series data as input and they directly optimize simulated soil water content by the 1-D Richardson–Richards equation with observations, ensuring improved applicability to field conditions. We trained and tested the model using two years of soil moisture observations from 1,181 sites in the International Soil Moisture Network. Evaluation using field data demonstrates that SPTFs achieve a Nash-Sutcliffe Efficiency of 0.65 and root mean squared error of 0.072 cm<sup>3</sup> cm<sup>−3</sup> in simulating soil water content at the depth of 0.05 m on the test set (<i>n</i> = 179), which is close to the values predicted by the inverse modeling method, while maintaining the computational efficiency of PTFs. This study highlights the promise of SPTFs as a robust parameterization framework for localized field applications.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"46 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145962156","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}
M. A. Chaudhry, S. Kiemle, A. Pohlmeier, R. Helmig, J. A. Huisman
Saline water evaporation from porous media is a key phenomenon in the terrestrial environment and is linked to problems such as soil salinization and weathering of building materials. Recent modeling studies suggest the development of local instabilities due to density differences during evaporation in case of saturated porous media with high permeability. To experimentally investigate this and improve our understanding of near surface solute accumulation, we performed evaporation experiments on two types of porous media (F36 and W3) with intrinsic permeabilities that differed by two orders of magnitude. Using magnetic resonance imaging (23Na-MRI), we monitored the development of solute accumulation and subsequent redistribution during evaporation under wicking conditions. The F36 sample showed an initial enrichment at the surface, but soon after a downwards moving plume developed that redistributed NaCl into the column. Average depth profiles of Na concentrations obtained from 3D imaging showed that the surface concentration reached only 2.5 mol L−1, well below the solubility limit. In contrast, the W3 sample with lower permeability showed enrichment in a shallow near-surface zone reaching a concentration of over 6 mol L−1. No fingering occurred although the mean evaporation rate was similar to that of F36 sand. Comparison of experimental results with numerical simulations using DuMux for both samples showed qualitative agreement between measured and modeled solute concentrations. This study experimentally confirms the importance of density-driven redistribution of solutes in case of evaporating saturated porous media, carrying implications for predicting evaporation rates and the time to start of salt crust formation.
{"title":"Non-Invasive Imaging of Solute Redistribution Below Evaporating Surfaces Using 23Na-MRI","authors":"M. A. Chaudhry, S. Kiemle, A. Pohlmeier, R. Helmig, J. A. Huisman","doi":"10.1029/2025wr041207","DOIUrl":"https://doi.org/10.1029/2025wr041207","url":null,"abstract":"Saline water evaporation from porous media is a key phenomenon in the terrestrial environment and is linked to problems such as soil salinization and weathering of building materials. Recent modeling studies suggest the development of local instabilities due to density differences during evaporation in case of saturated porous media with high permeability. To experimentally investigate this and improve our understanding of near surface solute accumulation, we performed evaporation experiments on two types of porous media (F36 and W3) with intrinsic permeabilities that differed by two orders of magnitude. Using magnetic resonance imaging (<sup>23</sup>Na-MRI), we monitored the development of solute accumulation and subsequent redistribution during evaporation under wicking conditions. The F36 sample showed an initial enrichment at the surface, but soon after a downwards moving plume developed that redistributed NaCl into the column. Average depth profiles of Na concentrations obtained from 3D imaging showed that the surface concentration reached only 2.5 mol L<sup>−1</sup>, well below the solubility limit. In contrast, the W3 sample with lower permeability showed enrichment in a shallow near-surface zone reaching a concentration of over 6 mol L<sup>−1</sup>. No fingering occurred although the mean evaporation rate was similar to that of F36 sand. Comparison of experimental results with numerical simulations using DuMu<sup>x</sup> for both samples showed qualitative agreement between measured and modeled solute concentrations. This study experimentally confirms the importance of density-driven redistribution of solutes in case of evaporating saturated porous media, carrying implications for predicting evaporation rates and the time to start of salt crust formation.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"18 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145956225","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}
The prediction of river planimetric evolution and related interactions with anthropic activities and public safety is one of the most critical aspects in the planning of a sustainable land-use. Since the beginning of the past century, a large number of theoretical and experimental studies have focused on the investigation of river meandering dynamics, coming to sometimes contrasting conclusions in the forecast of the associated bend sequence pattern. Drawing inspiration from the phenomenological equivalence between fluid-dynamic and morpho-dynamic dispersion within the river floodplain, the present contribution proposes an explicit analytical solution in terms of scale-dependent and equilibrium sinuosity. Such analytical solution, which reveals the strong dependence of river equilibrium planform on valley bank-full velocity distribution, is successfully validated on the basis of a field data set provided via a restoration pilot project by Basilicata Region Environment and Energy Department (Italy), and further discussed by related lagrangian simulations. Moreover, the governing equation from which the equilibrium solution originates is shown to be compatible with the interpretation of near-equilibrium dynamics highlighted by stochastic numerical experiments documented in the literature.
{"title":"A Cascade-Like Energy Dissipation Mechanism Behind the Gradual Achievement of River Equilibrium Sinuosity","authors":"M. Pannone","doi":"10.1029/2025wr041123","DOIUrl":"https://doi.org/10.1029/2025wr041123","url":null,"abstract":"The prediction of river planimetric evolution and related interactions with anthropic activities and public safety is one of the most critical aspects in the planning of a sustainable land-use. Since the beginning of the past century, a large number of theoretical and experimental studies have focused on the investigation of river meandering dynamics, coming to sometimes contrasting conclusions in the forecast of the associated bend sequence pattern. Drawing inspiration from the phenomenological equivalence between fluid-dynamic and morpho-dynamic dispersion within the river floodplain, the present contribution proposes an explicit analytical solution in terms of scale-dependent and equilibrium sinuosity. Such analytical solution, which reveals the strong dependence of river equilibrium planform on valley bank-full velocity distribution, is successfully validated on the basis of a field data set provided via a restoration pilot project by Basilicata Region Environment and Energy Department (Italy), and further discussed by related lagrangian simulations. Moreover, the governing equation from which the equilibrium solution originates is shown to be compatible with the interpretation of near-equilibrium dynamics highlighted by stochastic numerical experiments documented in the literature.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"49 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145956221","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}
Yubo Jia, Xiaoling Su, Vijay P. Singh, Bingnan Zhao, Te Zhang, Jiangdong Chu, Haijiang Wu
Accurate runoff forecasting helps mitigate flooding and drought risks and ensure water security under changing conditions. Compared to deterministic prediction models, interval prediction can more effectively quantify uncertainty, enhancing practical applicability. However, the Mixture Density Network (MDN) model—a state-of-the-art probabilistic modeling approach in hydrology—is susceptible to bias from distributional misspecification, and its prediction intervals are often overly wide, reducing practical utility. We therefore innovatively incorporated the Weighted Conformal Inference (WCI) strategy, which accounts for distributional shifts in runoff sequences, and integrated it with MDN to develop the WCI-MDN model for runoff interval prediction. To validate the effectiveness of the WCI strategy, we constructed six models in total: MDNs and WCI-MDNs under three distributions—Gaussian Mixture (GMM), Laplace Mixture (LMM), and Countable Mixtures of Asymmetric Laplacians (CMAL)—and evaluated their accuracy and robustness using data from 222 basins in the CAMELS-AUS data set. Results indicated that among the three MDN models, the LMM distribution achieved the best interval prediction performance, followed by the CMAL and GMM distributions. After introducing the WCI strategy, the coverage width-based criterion (CWC) for GMM, LMM, and CMAL distributions decreased by approximately 61.1%, 48.7%, and 54.3%, respectively, across all basins, demonstrating that the WCI-MDNs achieved higher prediction reliability. Furthermore, compared to the MDNs, the standard deviation of the CWC for the WCI-MDNs was reduced by 66.7%–81.8%, indicating higher robustness. Thus, the study improved the existing MDNs, providing a promising new approach for runoff interval prediction.
{"title":"A Novel Hybrid Predictive Model Based on Mixture Density Networks With Weighted Conformal Inference Strategy for Runoff Interval Prediction Across Australia","authors":"Yubo Jia, Xiaoling Su, Vijay P. Singh, Bingnan Zhao, Te Zhang, Jiangdong Chu, Haijiang Wu","doi":"10.1029/2024wr039807","DOIUrl":"https://doi.org/10.1029/2024wr039807","url":null,"abstract":"Accurate runoff forecasting helps mitigate flooding and drought risks and ensure water security under changing conditions. Compared to deterministic prediction models, interval prediction can more effectively quantify uncertainty, enhancing practical applicability. However, the Mixture Density Network (MDN) model—a state-of-the-art probabilistic modeling approach in hydrology—is susceptible to bias from distributional misspecification, and its prediction intervals are often overly wide, reducing practical utility. We therefore innovatively incorporated the Weighted Conformal Inference (WCI) strategy, which accounts for distributional shifts in runoff sequences, and integrated it with MDN to develop the WCI-MDN model for runoff interval prediction. To validate the effectiveness of the WCI strategy, we constructed six models in total: MDNs and WCI-MDNs under three distributions—Gaussian Mixture (GMM), Laplace Mixture (LMM), and Countable Mixtures of Asymmetric Laplacians (CMAL)—and evaluated their accuracy and robustness using data from 222 basins in the CAMELS-AUS data set. Results indicated that among the three MDN models, the LMM distribution achieved the best interval prediction performance, followed by the CMAL and GMM distributions. After introducing the WCI strategy, the coverage width-based criterion (CWC) for GMM, LMM, and CMAL distributions decreased by approximately 61.1%, 48.7%, and 54.3%, respectively, across all basins, demonstrating that the WCI-MDNs achieved higher prediction reliability. Furthermore, compared to the MDNs, the standard deviation of the CWC for the WCI-MDNs was reduced by 66.7%–81.8%, indicating higher robustness. Thus, the study improved the existing MDNs, providing a promising new approach for runoff interval prediction.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":"24 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145968874","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：