Pub Date : 2025-02-05DOI: 10.1016/j.jhydrol.2025.132815
Shenqiang Chen , Jeremy K.C. Rugenstein , Andreas Mulch
The Pamir range, located in Central Asia, mainly receives moisture from the mid-latitude westerlies, but its western side (i.e., Tajikistan Pamir) receives much of its precipitation in the winter and spring and its eastern side (i.e., Chinese Pamir) in the summer. Thus, the Pamir provides a natural laboratory to study the distribution of surface water stable isotopes across a large mountain range that ultimately receives moisture from one single source but has different precipitation seasonality regimes between its two sides. In this study, we present stable oxygen (δ18O) and hydrogen (δ2H) isotope data for 113 surface water samples from the Chinese Pamir. Our new data, along with previously published stable isotope data, show that the slope of the Chinese Pamir local meteoric water line is higher than that of the Global Meteoric Water Line (GMWL), and almost all of the data plot above the GMWL, implying that the Chinese Pamir surface waters have not experienced significant isotopic modification by evaporation. The Chinese Pamir surface waters have substantially higher δ18O and d-excess values and a steeper apparent δ18O lapse rate than surface water samples collected from the Tajikistan Pamir. We suggest that this contrast results from the shift in precipitation seasonality across the Pamir, with dominantly winter and springtime precipitation on the Tajikistan side and summertime precipitation on the Chinese side of the Pamir. This predominant summertime precipitation regime results in surface waters with high δ18O values in the Chinese Pamir. Further, this summertime moisture is dominantly convectively recycled moisture, resulting in high d-excess values in surface waters. The percentage of summertime moisture, which has high δ18O values, decreases west and with elevation in the Chinese Pamir, resulting in a steep apparent δ18O lapse rate of − 3.2 ‰/km. The importance of precipitation seasonality in modulating δ18O values across the Pamir suggests that proxy-derived records of past environments in the region must consider the mechanisms that today cause the seasonality contrast.
{"title":"Stable isotope composition of surface waters across the Pamir, Central Asia: Implications of precipitation seasonality","authors":"Shenqiang Chen , Jeremy K.C. Rugenstein , Andreas Mulch","doi":"10.1016/j.jhydrol.2025.132815","DOIUrl":"10.1016/j.jhydrol.2025.132815","url":null,"abstract":"<div><div>The Pamir range, located in Central Asia, mainly receives moisture from the mid-latitude westerlies, but its western side (i.e., Tajikistan Pamir) receives much of its precipitation in the winter and spring and its eastern side (i.e., Chinese Pamir) in the summer. Thus, the Pamir provides a natural laboratory to study the distribution of surface water stable isotopes across a large mountain range that ultimately receives moisture from one single source but has different precipitation seasonality regimes between its two sides. In this study, we present stable oxygen (δ<sup>18</sup>O) and hydrogen (δ<sup>2</sup>H) isotope data for 113 surface water samples from the Chinese Pamir. Our new data, along with previously published stable isotope data, show that the slope of the Chinese Pamir local meteoric water line is higher than that of the Global Meteoric Water Line (GMWL), and almost all of the data plot above the GMWL, implying that the Chinese Pamir surface waters have not experienced significant isotopic modification by evaporation. The Chinese Pamir surface waters have substantially higher δ<sup>18</sup>O and d-excess values and a steeper apparent δ<sup>18</sup>O lapse rate than surface water samples collected from the Tajikistan Pamir. We suggest that this contrast results from the shift in precipitation seasonality across the Pamir, with dominantly winter and springtime precipitation on the Tajikistan side and summertime precipitation on the Chinese side of the Pamir. This predominant summertime precipitation regime results in surface waters with high δ<sup>18</sup>O values in the Chinese Pamir. Further, this summertime moisture is dominantly convectively recycled moisture, resulting in high d-excess values in surface waters. The percentage of summertime moisture, which has high δ<sup>18</sup>O values, decreases west and with elevation in the Chinese Pamir, resulting in a steep apparent δ<sup>18</sup>O lapse rate of − 3.2 ‰/km. The importance of precipitation seasonality in modulating δ<sup>18</sup>O values across the Pamir suggests that proxy-derived records of past environments in the region must consider the mechanisms that today cause the seasonality contrast.</div></div>","PeriodicalId":362,"journal":{"name":"Journal of Hydrology","volume":"653 ","pages":"Article 132815"},"PeriodicalIF":5.9,"publicationDate":"2025-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143285942","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}
California is highly vulnerable to extreme precipitation events due to the dense landfall of atmospheric rivers (ARs) during the winter months, often resulting in catastrophic consequences such as widespread floods, mudslides, and landslides. This study focuses on the recovery of daily variations in AR-driven terrestrial water storage (TWS), which produces geodetically detectable ground subsidence. We invert GNSS vertical positions to obtain daily estimates of equivalent water height (EWH) through a variational Bayesian principal component analysis (vbPCA) based inversion scheme and track significant water gains during record-setting winter months in four water years (WYs) 2011, 2017, 2019, and 2023. These precipitation extremes have resulted in a substantial short-term increase in water storage, as evidenced by the multi-source EWH datasets (GNSS, GRACE, and NLDAS). Notably, WY 2023 experienced the highest snowfall due to the landfalls of high-density, high-category ARs, while WY 2017 recorded the highest precipitation totals, driven by the most frequent occurrence of hazardous ARs. Our findings further highlight that GNSS can accurately detect exceptionally wet hydrological events on short time scales, benefiting from an improved signal-to-noise ratio due to substantial increase in water storage. The results also indicate that while these extreme water years can help alleviate surface subsidence in the Central Valley caused by groundwater overexploitation, it is insufficient to alter California’s heavy reliance on groundwater for its intensive agricultural activities. Our findings demonstrate that GNSS is successful in tracking prodigious water increases from short-term precipitation extremes that are weaker than powerful hurricanes, illuminating the prospect of GNSS in supporting water management and flood preparedness.
{"title":"Tracking California’s striking water storage gains attributed to intensive atmospheric rivers","authors":"Zhongshan Jiang , Hui Zhang , Miao Tang , Xinghai Yang , Linguo Yuan , Yuan Yuan , Wei Feng , Min Zhong","doi":"10.1016/j.jhydrol.2025.132804","DOIUrl":"10.1016/j.jhydrol.2025.132804","url":null,"abstract":"<div><div>California is highly vulnerable to extreme precipitation events due to the dense landfall of atmospheric rivers (ARs) during the winter months, often resulting in catastrophic consequences such as widespread floods, mudslides, and landslides. This study focuses on the recovery of daily variations in AR-driven terrestrial water storage (TWS), which produces geodetically detectable ground subsidence. We invert GNSS vertical positions to obtain daily estimates of equivalent water height (EWH) through a variational Bayesian principal component analysis (vbPCA) based inversion scheme and track significant water gains during record-setting winter months in four water years (WYs) 2011, 2017, 2019, and 2023. These precipitation extremes have resulted in a substantial short-term increase in water storage, as evidenced by the multi-source EWH datasets (GNSS, GRACE, and NLDAS). Notably, WY 2023 experienced the highest snowfall due to the landfalls of high-density, high-category ARs, while WY 2017 recorded the highest precipitation totals, driven by the most frequent occurrence of hazardous ARs. Our findings further highlight that GNSS can accurately detect exceptionally wet hydrological events on short time scales, benefiting from an improved signal-to-noise ratio due to substantial increase in water storage. The results also indicate that while these extreme water years can help alleviate surface subsidence in the Central Valley caused by groundwater overexploitation, it is insufficient to alter California’s heavy reliance on groundwater for its intensive agricultural activities. Our findings demonstrate that GNSS is successful in tracking prodigious water increases from short-term precipitation extremes that are weaker than powerful hurricanes, illuminating the prospect of GNSS in supporting water management and flood preparedness.</div></div>","PeriodicalId":362,"journal":{"name":"Journal of Hydrology","volume":"653 ","pages":"Article 132804"},"PeriodicalIF":5.9,"publicationDate":"2025-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143372484","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}
Pub Date : 2025-02-04DOI: 10.1016/j.jhydrol.2025.132753
Zidong Pan , Zhilin Guo , Kewei Chen , Wenxi Lu , Chunmiao Zheng
<div><div>Groundwater contaminant source estimation (GCSE) plays a vital role in the risk assessment and remediation of groundwater contamination. GCSE involves determining the optimal values of unknown variables that result in the observed contaminant concentrations at monitoring wells. This can be achieved by establishing an inverse mapping from the observed concentrations to the unknown variables characterizing the contaminant source. In recent years, deep learning such as generative adversarial neural network (GAN) has gained increasing attention as an effective tool for establishing inverse mapping relationships between observed contaminant distributions and source-related parameters. While a single-directional GAN can effectively establish this inverse mapping relationship, it faces a significant limitation in that the inverse results cannot be validated by comparing the corresponding simulation outputs with the observed contaminant concentrations, which is crucial for ensuring accuracy and reliability in real-world GCSE applications. To address this issue, we propose a bidirectional generative adversarial neural network (Bi-GAN) that incorporates both an inversion process and a forward process, enhancing the supervision of the inverse mapping. Specifically, the inversion process produces unknown variables estimates (generated data) from the observed contaminant concentrations, while the forward process converts these estimates into the corresponding simulation outputs. The forward process then evaluates the similarity between the simulation outputs and observed concentrations. This similarity measure informs the training of the inversion process, ensuring greater accuracy. Once the model meets the accuracy threshold, the inversion process can be extracted for providing reliable GCSE estimations. In addition, the performance of Bi-GAN is strongly influenced by the quality of its training samples. To optimize this, we introduced an adaptive sampling strategy, which significantly improves the quality of the training data, resulting in enhanced accuracy for GCSE (50 % error reduction in case 1 and 70 % error reduction in case 2). Furthermore, the data-driven nature of the Bi-GAN allows for a substantial reduction in inversion inference time cost during the estimation process. The proposed Bi-GAN framework was evaluated using two hypothetical cases: a heterogeneity with zone partitioning case (case 1) and a heterogeneity with continuous medium case (case 2), aiming to provide an efficient, accurate, and cost-effective solution for GCSE. The results demonstrate that Bi-GAN delivers superior performance, achieving high estimation accuracy (0.87 % ARE in case 1 and 3.62 % ARE in case 2) and remarkable computational efficiency (0.05 s and 0.07 s inversion inference time). These results are particularly noteworthy when compared with traditional inversion methods in GCSE such as the ensemble kalman filter (EnKF) and the genetic algorithm (GA).</div><
{"title":"A deep adaptive bidirectional generative adversarial neural network (Bi-GAN) for groundwater contamination source estimation","authors":"Zidong Pan , Zhilin Guo , Kewei Chen , Wenxi Lu , Chunmiao Zheng","doi":"10.1016/j.jhydrol.2025.132753","DOIUrl":"10.1016/j.jhydrol.2025.132753","url":null,"abstract":"<div><div>Groundwater contaminant source estimation (GCSE) plays a vital role in the risk assessment and remediation of groundwater contamination. GCSE involves determining the optimal values of unknown variables that result in the observed contaminant concentrations at monitoring wells. This can be achieved by establishing an inverse mapping from the observed concentrations to the unknown variables characterizing the contaminant source. In recent years, deep learning such as generative adversarial neural network (GAN) has gained increasing attention as an effective tool for establishing inverse mapping relationships between observed contaminant distributions and source-related parameters. While a single-directional GAN can effectively establish this inverse mapping relationship, it faces a significant limitation in that the inverse results cannot be validated by comparing the corresponding simulation outputs with the observed contaminant concentrations, which is crucial for ensuring accuracy and reliability in real-world GCSE applications. To address this issue, we propose a bidirectional generative adversarial neural network (Bi-GAN) that incorporates both an inversion process and a forward process, enhancing the supervision of the inverse mapping. Specifically, the inversion process produces unknown variables estimates (generated data) from the observed contaminant concentrations, while the forward process converts these estimates into the corresponding simulation outputs. The forward process then evaluates the similarity between the simulation outputs and observed concentrations. This similarity measure informs the training of the inversion process, ensuring greater accuracy. Once the model meets the accuracy threshold, the inversion process can be extracted for providing reliable GCSE estimations. In addition, the performance of Bi-GAN is strongly influenced by the quality of its training samples. To optimize this, we introduced an adaptive sampling strategy, which significantly improves the quality of the training data, resulting in enhanced accuracy for GCSE (50 % error reduction in case 1 and 70 % error reduction in case 2). Furthermore, the data-driven nature of the Bi-GAN allows for a substantial reduction in inversion inference time cost during the estimation process. The proposed Bi-GAN framework was evaluated using two hypothetical cases: a heterogeneity with zone partitioning case (case 1) and a heterogeneity with continuous medium case (case 2), aiming to provide an efficient, accurate, and cost-effective solution for GCSE. The results demonstrate that Bi-GAN delivers superior performance, achieving high estimation accuracy (0.87 % ARE in case 1 and 3.62 % ARE in case 2) and remarkable computational efficiency (0.05 s and 0.07 s inversion inference time). These results are particularly noteworthy when compared with traditional inversion methods in GCSE such as the ensemble kalman filter (EnKF) and the genetic algorithm (GA).</div><","PeriodicalId":362,"journal":{"name":"Journal of Hydrology","volume":"653 ","pages":"Article 132753"},"PeriodicalIF":5.9,"publicationDate":"2025-02-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143350710","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}
In this study, H2018 function, which was proposed as a suitable function for simulating Flow Duration Curves (FDC) in the previous studies, was fitted to 1915 hydrometric gauges across the United States. The observed streamflow for all these gauges during the 1984–2016 period were partitioned into three intervals (1984–1994, 1995–2005, and 2006–2016) and separate Random Forest (RF) models were trained to estimate H2018 function parameters for each time interval. For this purpose, 85 basin characteristics (representing hydrology, morphology, topography, land cover, soil, dam, and climate) were used as RF model predictors. RF feature importance results showed that Base Flow Index is an important feature for estimating three out of four parameters of the H2018 function. Comparison between feature importance values in the three time periods showed increasing influence of climate characteristics on FDC in the latest period (2006–2016) highlighting the importance of climate change impacts. Next, features with lower importance values were removed from the predictor set of the RF models. The results showed that reducing RF model features from 85 to 16 did not negatively impact the performance of the models making the generalization of the proposed methodology easier. Finally, the streamflow-related RF predictor (BFI) was removed, and the RF models with 15 input features were recalibrated to be used for the remaining 1282 pseudo-ungauged basins assuming no hydrometric station exists. Overall, it was concluded that the proposed methodology can be effectively used for estimation of FDC curves for pseudo-ungauged basins with Nash-Sutcliffe values of about 0.98. Diversity of the climatic and hydrologic conditions of the basins used in this study, variety of basin characteristics used as RF model input features, explaining how they can be efficiently reduced to be applicable to pseudo-ungauged basins and presenting how the importance of these features has changed over time were the main novel contributions and findings of this study.
{"title":"Estimation of standardized flow Duration curve for gauged and ungauged basins","authors":"Pegah Palizban , Banafsheh Zahraie , Neda Dolatabadi","doi":"10.1016/j.jhydrol.2025.132787","DOIUrl":"10.1016/j.jhydrol.2025.132787","url":null,"abstract":"<div><div>In this study, H2018 function, which was proposed as a suitable function for simulating Flow Duration Curves (FDC) in the previous studies, was fitted to 1915 hydrometric gauges across the United States. The observed streamflow for all these gauges during the 1984–2016 period were partitioned into three intervals (1984–1994, 1995–2005, and 2006–2016) and separate Random Forest (RF) models were trained to estimate H2018 function parameters for each time interval. For this purpose, 85 basin characteristics (representing hydrology, morphology, topography, land cover, soil, dam, and climate) were used as RF model predictors. RF feature importance results showed that Base Flow Index is an important feature for estimating three out of four parameters of the H2018 function. Comparison between feature importance values in the three time periods showed increasing influence of climate characteristics on FDC in the latest period (2006–2016) highlighting the importance of climate change impacts. Next, features with lower importance values were removed from the predictor set of the RF models. The results showed that reducing RF model features from 85 to 16 did not negatively impact the performance of the models making the generalization of the proposed methodology easier. Finally, the streamflow-related RF predictor (BFI) was removed, and the RF models with 15 input features were recalibrated to be used for the remaining 1282 pseudo-ungauged basins assuming no hydrometric station exists. Overall, it was concluded that the proposed methodology can be effectively used for estimation of FDC curves for pseudo-ungauged basins with Nash-Sutcliffe values of about 0.98. Diversity of the climatic and hydrologic conditions of the basins used in this study, variety of basin characteristics used as RF model input features, explaining how they can be efficiently reduced to be applicable to pseudo-ungauged basins and presenting how the importance of these features has changed over time were the main novel contributions and findings of this study.</div></div>","PeriodicalId":362,"journal":{"name":"Journal of Hydrology","volume":"653 ","pages":"Article 132787"},"PeriodicalIF":5.9,"publicationDate":"2025-02-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143372483","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}
Pub Date : 2025-02-04DOI: 10.1016/j.jhydrol.2025.132800
Nie Zhou , Hua Chen , Chong-Yu Xu , Bingyi Liu , Jing Yang
Space-time image velocimetry (STIV) is a video-based technique for measuring river surface flow velocities and is widely used owing to its simplicity, efficiency, and safety. However, a key limitation of traditional STIV is its reliance on a preset velocity measurement line, which can lead to significant errors when the actual flow direction deviates from this predefined line. To overcome this limitation, a novel adaptive flow direction search algorithm based on Hough transform is proposed. The algorithm dynamically adjusts the measurement line direction in real time based on the flow conditions at the measurement point, thereby enabling more accurate surface flow velocity measurements. The proposed method was verified using synthetic videos, and its applicability for surface flow velocity was confirmed through comparisons with the traditional fixed-direction method and an acoustic Doppler current profiler (ADCP) both in artificial channels and natural rivers. The results revealed that the proposed method effectively identified and adapted to different flow directions in synthetic videos with a mean absolute error less than 1.0°. Furthermore, with the adaptive flow direction search algorithm, the measurement accuracy of the surface flow velocity estimation was significantly improved compared with traditional fixed direction method, with the relative error in the cross-sectional average flow velocity being less than 5 % under both artificial channel and natural river conditions. The proposed adaptive flow direction search algorithm can help in enhancing the stability of the STIV technique under complex flow conditions, and provides more precise and reliable river surface flow velocity measurement.
{"title":"Improving river surface flow velocity measurement by coupling adaptive flow direction search algorithm with space-time image velocimetry","authors":"Nie Zhou , Hua Chen , Chong-Yu Xu , Bingyi Liu , Jing Yang","doi":"10.1016/j.jhydrol.2025.132800","DOIUrl":"10.1016/j.jhydrol.2025.132800","url":null,"abstract":"<div><div>Space-time image velocimetry (STIV) is a video-based technique for measuring river surface flow velocities and is widely used owing to its simplicity, efficiency, and safety. However, a key limitation of traditional STIV is its reliance on a preset velocity measurement line, which can lead to significant errors when the actual flow direction deviates from this predefined line. To overcome this limitation, a novel adaptive flow direction search algorithm based on Hough transform is proposed. The algorithm dynamically adjusts the measurement line direction in real time based on the flow conditions at the measurement point, thereby enabling more accurate surface flow velocity measurements. The proposed method was verified using synthetic videos, and its applicability for surface flow velocity was confirmed through comparisons with the traditional fixed-direction method and an acoustic Doppler current profiler (ADCP) both in artificial channels and natural rivers. The results revealed that the proposed method effectively identified and adapted to different flow directions in synthetic videos with a mean absolute error less than 1.0°. Furthermore, with the adaptive flow direction search algorithm, the measurement accuracy of the surface flow velocity estimation was significantly improved compared with traditional fixed direction method, with the relative error in the cross-sectional average flow velocity being less than 5 % under both artificial channel and natural river conditions. The proposed adaptive flow direction search algorithm can help in enhancing the stability of the STIV technique under complex flow conditions, and provides more precise and reliable river surface flow velocity measurement.</div></div>","PeriodicalId":362,"journal":{"name":"Journal of Hydrology","volume":"653 ","pages":"Article 132800"},"PeriodicalIF":5.9,"publicationDate":"2025-02-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143286004","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}
Pub Date : 2025-02-03DOI: 10.1016/j.jhydrol.2025.132732
Haowen Xie , Yawen Wu , Sylvana Melo dos Santos
<div><div>Green roofs (GRs) have gained global attention and promotion for their social and environmental benefits. However, their limited capacity to retain rainwater leads to water runoff and wastage. Enhancing the collection and utilization capacity of rainwater from GRs is a current research trend. This paper aims to explore the potential of adding rooftop rainwater harvesting (RWH) systems to GRs as a solution to address this issue.</div><div>Currently, there is limited research in green roof rainwater harvesting system (GR-RWH), with a primary focus on the water reliability of GR-RWH with fixed design parameters, such as roof area and tank volume. Key questions, such as how the optimal tank volume varies with different roof area, water demand, and substrate depth, and the differences in investment payback period and water reliability of GR-RWH systems under varying roof area, water demand, tank volume, and substrate depth, require urgent investigation.</div><div>This paper proposes a modified selfish herd optimization algorithm (MSHO) by integrating the Beetle Antennae search strategy (BASS) into the traditional selfish herd optimization (SHO) algorithm. The integration of BASS is anticipated to enhance the MSHO algorithm’s ability to evade local optima and converge more rapidly to the global optimum in global searches, by simulating the dynamic search behavior of beetle antennae. Using MSHO, the paper conducts optimization calculations for multiple scenarios (3 substrate depths, 11 roof areas, and 11 water demands) to determine the optimal tank volume while considering water reliability and investment payback period as decision variables. The corresponding water reliability and investment payback period are obtained for each scenario.</div><div>The paper concludes the following: (1) Increasing substrate depth leads to a continuous decrease in the maximum water demand that the GR-RWH system can meet, while tank volume increases, water reliability slightly decreases, and investment payback period increases. (2) Increasing roof area results in a higher available rainwater volume, allowing for higher water demand and water reliability, and a reduction in investment payback period. However, for roof area values where water demand ranges from 0.1 to 0.5 m<sup>3</sup>/d, water reliability approaches 1.0, but investment payback period becomes excessively long. (3) For GRs with substrate depth of 50 mm, water demand greater than 0.5 m<sup>3</sup>/d, and roof area larger than 400 m<sup>2</sup>, it is recommended to add a RWH system. The optimal tank volume ranges from 5.5 to 85.5 m<sup>3</sup> for these scenarios, with a maximum water reliability of 0.5 to 1.0 and a minimum investment payback period of 11 to 38 years.</div><div>The optimization method presented in this study offers not only a novel theoretical perspective but also holds significant potential value for urban water resource management, particularly in the control of urban non-point so
{"title":"Optimizing green roof rainwater harvesting systems: A modified selfish Herds algorithm approach","authors":"Haowen Xie , Yawen Wu , Sylvana Melo dos Santos","doi":"10.1016/j.jhydrol.2025.132732","DOIUrl":"10.1016/j.jhydrol.2025.132732","url":null,"abstract":"<div><div>Green roofs (GRs) have gained global attention and promotion for their social and environmental benefits. However, their limited capacity to retain rainwater leads to water runoff and wastage. Enhancing the collection and utilization capacity of rainwater from GRs is a current research trend. This paper aims to explore the potential of adding rooftop rainwater harvesting (RWH) systems to GRs as a solution to address this issue.</div><div>Currently, there is limited research in green roof rainwater harvesting system (GR-RWH), with a primary focus on the water reliability of GR-RWH with fixed design parameters, such as roof area and tank volume. Key questions, such as how the optimal tank volume varies with different roof area, water demand, and substrate depth, and the differences in investment payback period and water reliability of GR-RWH systems under varying roof area, water demand, tank volume, and substrate depth, require urgent investigation.</div><div>This paper proposes a modified selfish herd optimization algorithm (MSHO) by integrating the Beetle Antennae search strategy (BASS) into the traditional selfish herd optimization (SHO) algorithm. The integration of BASS is anticipated to enhance the MSHO algorithm’s ability to evade local optima and converge more rapidly to the global optimum in global searches, by simulating the dynamic search behavior of beetle antennae. Using MSHO, the paper conducts optimization calculations for multiple scenarios (3 substrate depths, 11 roof areas, and 11 water demands) to determine the optimal tank volume while considering water reliability and investment payback period as decision variables. The corresponding water reliability and investment payback period are obtained for each scenario.</div><div>The paper concludes the following: (1) Increasing substrate depth leads to a continuous decrease in the maximum water demand that the GR-RWH system can meet, while tank volume increases, water reliability slightly decreases, and investment payback period increases. (2) Increasing roof area results in a higher available rainwater volume, allowing for higher water demand and water reliability, and a reduction in investment payback period. However, for roof area values where water demand ranges from 0.1 to 0.5 m<sup>3</sup>/d, water reliability approaches 1.0, but investment payback period becomes excessively long. (3) For GRs with substrate depth of 50 mm, water demand greater than 0.5 m<sup>3</sup>/d, and roof area larger than 400 m<sup>2</sup>, it is recommended to add a RWH system. The optimal tank volume ranges from 5.5 to 85.5 m<sup>3</sup> for these scenarios, with a maximum water reliability of 0.5 to 1.0 and a minimum investment payback period of 11 to 38 years.</div><div>The optimization method presented in this study offers not only a novel theoretical perspective but also holds significant potential value for urban water resource management, particularly in the control of urban non-point so","PeriodicalId":362,"journal":{"name":"Journal of Hydrology","volume":"653 ","pages":"Article 132732"},"PeriodicalIF":5.9,"publicationDate":"2025-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143349394","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}
Pub Date : 2025-02-03DOI: 10.1016/j.jhydrol.2025.132775
M.D. Fidelibus , G. Balacco , M.R. Alfio , M. Arfaoui , D. Bassukas , C. Güler , F. Hamzaoui-Azaza , C. Külls , A. Panagopoulos , A. Parisi , E. Sachsamanoglou , E. Tziritis
Seawater intrusion is the primary cause of groundwater salinisation in coastal aquifers. However, attributing salinisation solely to seawater intrusion may not always be accurate, given the likely presence of other sources. To understand if salinisation comes from seawater intrusion and its onset is crucial for groundwater management, but there are no definite threshold values for common indicators such as chlorides. Based on 1662 groundwater analyses from five Mediterranean coastal aquifers, the study aimed to distinguish the effects of mixing with present-day seawater from those caused by other sources. The trend analysis of cumulative probability plots of chloride (and total dissolved solids) is a key method for discriminating different groundwater salinisation sources and processes. Results establish that chloride, as a non-reactive tracer, is a more reliable indicator of seawater intrusion than total dissolved solids, a reactive indicator.
A chloride concentration threshold of 200 mg/L identifies the seawater intrusion onset. The threshold validation comes from groundwater salinisation facies, as provided by groundwater-type codification.
Fresh groundwater (Cl < 200 mg/L) anomalous total dissolved solids highlight the input of non-chloride salts and pollutants, providing caution regarding using total dissolved solids to recognise seawater intrusion. Beyond the threshold (Cl > 200 mg/L), data disclose emergent signals of salinisation sources and water–rock interaction processes overlapping seawater intrusion or the involvement of saline fluids different from present-day seawater. The threshold and a new categorisation of groundwater in coastal aquifers according to salinisation processes provide a benchmark for identifying and managing seawater intrusion in the Mediterranean area.
{"title":"A chloride threshold to identify the onset of seawater/saltwater intrusion and a novel categorization of groundwater in coastal aquifers","authors":"M.D. Fidelibus , G. Balacco , M.R. Alfio , M. Arfaoui , D. Bassukas , C. Güler , F. Hamzaoui-Azaza , C. Külls , A. Panagopoulos , A. Parisi , E. Sachsamanoglou , E. Tziritis","doi":"10.1016/j.jhydrol.2025.132775","DOIUrl":"10.1016/j.jhydrol.2025.132775","url":null,"abstract":"<div><div>Seawater intrusion is the primary cause of groundwater salinisation in coastal aquifers. However, attributing salinisation solely to seawater intrusion may not always be accurate, given the likely presence of other sources. To understand if salinisation comes from seawater intrusion and its onset is crucial for groundwater management, but there are no definite threshold values for common indicators such as chlorides. Based on 1662 groundwater analyses from five Mediterranean coastal aquifers, the study aimed to distinguish the effects of mixing with present-day seawater from those caused by other sources. The trend analysis of cumulative probability plots of chloride (and total dissolved solids) is a key method for discriminating different groundwater salinisation sources and processes. Results establish that chloride, as a non-reactive tracer, is a more reliable indicator of seawater intrusion than total dissolved solids, a reactive indicator.</div><div>A chloride concentration threshold of 200 mg/L identifies the seawater intrusion onset. The threshold validation comes from groundwater salinisation facies, as provided by groundwater-type codification.</div><div>Fresh groundwater (Cl < 200 mg/L) anomalous total dissolved solids highlight the input of non-chloride salts and pollutants, providing caution regarding using total dissolved solids to recognise seawater intrusion. Beyond the threshold (Cl > 200 mg/L), data disclose emergent signals of salinisation sources and water–rock interaction processes overlapping seawater intrusion or the involvement of saline fluids different from present-day seawater. The threshold and a new categorisation of groundwater in coastal aquifers according to salinisation processes provide a benchmark for identifying and managing seawater intrusion in the Mediterranean area.</div></div>","PeriodicalId":362,"journal":{"name":"Journal of Hydrology","volume":"653 ","pages":"Article 132775"},"PeriodicalIF":5.9,"publicationDate":"2025-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143175208","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-02DOI: 10.1016/j.jhydrol.2025.132773
Suelen Matiasso Fachi , Jirka Šimůnek , Rodrigo Pivoto Mulazzani , Quirijn De Jong Van Lier , Dalvan José Reinert , Paulo Ivonir Gubiani
Stony soils represent a complex porous system, which makes the description of their water retention a challenge. This study aims to assess whether the description of water retention in stony soils requires a bimodal Water Retention Curve (WRC) and to test the corresponding soil hydraulic parameters in Hydrus-1D for simulating drainage experiments. For this, we conducted field drainage experiments on stony soil profiles with different granulometric distributions and different amounts of rock fragments. The soil water content (θ) and the matric potential (h) were monitored at different soil depths during field drainage (h ≥ −100 cm) and evaporation (h < −100 cm) experiments, as well as determined in the laboratory for a drier range (h < −5000 cm) with a dew point psychrometer on soil samples. The bimodal van Genuchten-Mualem function was evaluated using the RETC curve fitting software. Subsequently, we tested and, when necessary, optimized the soil hydraulic parameters in Hydrus-1D to simulate drainage experiments. Our results showed that a bimodal WRC is needed to describe water retention in soils with rock fragments. The methodology used in this study, which combined measurements of θ and h in the field during a drainage experiment followed by an evaporation experiment with the laboratory data with a dew point psychrometer, proved to be suitable in detecting the bimodal behavior of WRC of stony soils. The optimized soil hydraulic parameters were physically consistent, except for some high values of the retention curve parameters. Only a few parameters had to be optimized to simulate drainage with Hydrus-1D accurately. The Root Mean Square Error of the fitted WRCs ranged between 7.10−5 and 4.52.10−3 cm3 cm−3. The Root Mean Square Weighted Error (dimensionless) for the Hydrus-1D drainage simulations varied from 0.14 to 0.88 before parameter optimization and it was reduced by up to 12 times after parameter optimization.We conclude that the methodological strategy used in this study effectively detects the bimodal nature of WRC of soils with rock fragments and may be helpful for future studies in such conditions.
{"title":"Assessing bimodal water retention functions of stony soils using curve fitting and inverse modeling","authors":"Suelen Matiasso Fachi , Jirka Šimůnek , Rodrigo Pivoto Mulazzani , Quirijn De Jong Van Lier , Dalvan José Reinert , Paulo Ivonir Gubiani","doi":"10.1016/j.jhydrol.2025.132773","DOIUrl":"10.1016/j.jhydrol.2025.132773","url":null,"abstract":"<div><div>Stony soils represent a complex porous system, which makes the description of their water retention a challenge. This study aims to assess whether the description of water retention in stony soils requires a bimodal Water Retention Curve (WRC) and to test the corresponding soil hydraulic parameters in Hydrus-1D for simulating drainage experiments. For this, we conducted field drainage experiments on stony soil profiles with different granulometric distributions and different amounts of rock fragments. The soil water content (θ) and the matric potential (h) were monitored at different soil depths during field drainage (h ≥ −100 cm) and evaporation (h < −100 cm) experiments, as well as determined in the laboratory for a drier range (h < −5000 cm) with a dew point psychrometer on soil samples. The bimodal van Genuchten-Mualem function was evaluated using the RETC curve fitting software. Subsequently, we tested and, when necessary, optimized the soil hydraulic parameters in Hydrus-1D to simulate drainage experiments. Our results showed that a bimodal WRC is needed to describe water retention in soils with rock fragments. The methodology used in this study, which combined measurements of θ and h in the field during a drainage experiment followed by an evaporation experiment with the laboratory data with a dew point psychrometer, proved to be suitable in detecting the bimodal behavior of WRC of stony soils. The optimized soil hydraulic parameters were physically consistent, except for some high values of the retention curve parameters. Only a few parameters had to be optimized to simulate drainage with Hydrus-1D accurately. The Root Mean Square Error of the fitted WRCs ranged between 7.10<sup>−</sup><sup>5</sup> and 4.52.10<sup>−</sup><sup>3</sup> cm<sup>3</sup> cm<sup>−3</sup>. The Root Mean Square Weighted Error (dimensionless) for the Hydrus-1D drainage simulations varied from 0.14 to 0.88 before parameter optimization and it was reduced by up to 12 times after parameter optimization.We conclude that the methodological strategy used in this study effectively detects the bimodal nature of WRC of soils with rock fragments and may be helpful for future studies in such conditions.</div></div>","PeriodicalId":362,"journal":{"name":"Journal of Hydrology","volume":"654 ","pages":"Article 132773"},"PeriodicalIF":5.9,"publicationDate":"2025-02-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143419512","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}
Pub Date : 2025-02-02DOI: 10.1016/j.jhydrol.2025.132801
Faisal Baig , Luqman Ali , Muhammad Abrar Faiz , Haonan Chen , Mohsen Sherif
This paper presents a comprehensive approach to refine satellite precipitation estimates over the mountainous regions of the United Arab Emirates (UAE) using Long Short-Term Memory (LSTM). The primary aim is to address and correct biases in the CMORPH and IMERG satellite precipitation products by incorporating elevation, minimum temperature, and distance to coast covariates. The study is relying upon the acknowledgement of the substantial inconsistencies that often exist between in-situ gauge data and satellite-derived approximations, particularly in complex terrain areas where orographic effects marginally dominate rainfall trends. LSTM, known for its unique potential to process timeseries data and capture long-term dependencies, is employed to model the spatio-temporal dynamics of precipitation over the UAE’s mountainous regions. The study uses daily rainfall data for the duration 2004–2021 in addition to other topographical and climatological variables. The LSTM model was trained with these variables to identify the inherent biases in the CMORPH and IMERG relative to gauge observations. The analysis exhibits that the LSTM-based framework substantially improves the accuracy of precipitation products. The integration of terrain and climatic covariates not only aids in capturing the orographic enhancement of precipitation but also facilitates a more pronounced correction of biases, leading to improved precipitation data quality. The study employs statistical metrics such as the Nash-Sutcliffe Efficiency (NSE), Root Mean Square Error (RMSE), Normalized Mean Absolute Error (NMAE), Probability of Detection (POD), and Critical Success Index (CSI) to quantify the improvements achieved through the LSTM-based bias correction. NSE score was enhanced by almost 45 % for most of the station while the RMSE decreased in the range from 3.4 mm to 0.5 mm. Moreover, corrected products could capture the true rainfall with a 30 % higher accuracy than the raw products as shown by POD and CSI calculations. This study contributes to the field of hydrology and climate science by improving the accuracy of satellite-derived precipitation estimates in mountainous regions and sets a precedent for the application of advanced machine learning techniques in environmental science. The implications of this research are far-reaching, offering valuable insights for water resource management, agricultural planning, and disaster mitigation efforts in the UAE and similar regions worldwide.
{"title":"From bias to accuracy: Transforming satellite precipitation data in arid regions with machine learning and topographical insights","authors":"Faisal Baig , Luqman Ali , Muhammad Abrar Faiz , Haonan Chen , Mohsen Sherif","doi":"10.1016/j.jhydrol.2025.132801","DOIUrl":"10.1016/j.jhydrol.2025.132801","url":null,"abstract":"<div><div>This paper presents a comprehensive approach to refine satellite precipitation estimates over the mountainous regions of the United Arab Emirates (UAE) using Long Short-Term Memory (LSTM). The primary aim is to address and correct biases in the CMORPH and IMERG satellite precipitation products by incorporating elevation, minimum temperature, and distance to coast covariates. The study is relying upon the acknowledgement of the substantial inconsistencies that often exist between in-situ gauge data and satellite-derived approximations, particularly in complex terrain areas where orographic effects marginally dominate rainfall trends. LSTM, known for its unique potential to process timeseries data and capture long-term dependencies, is employed to model the spatio-temporal dynamics of precipitation over the UAE’s mountainous regions. The study uses daily rainfall data for the duration 2004–2021 in addition to other topographical and climatological variables. The LSTM model was trained with these variables to identify the inherent biases in the CMORPH and IMERG relative to gauge observations. The analysis exhibits that the LSTM-based framework substantially improves the accuracy of precipitation products. The integration of terrain and climatic covariates not only aids in capturing the orographic enhancement of precipitation but also facilitates a more pronounced correction of biases, leading to improved precipitation data quality. The study employs statistical metrics such as the Nash-Sutcliffe Efficiency (NSE), Root Mean Square Error (RMSE), Normalized Mean Absolute Error (NMAE), Probability of Detection (POD), and Critical Success Index (CSI) to quantify the improvements achieved through the LSTM-based bias correction. NSE score was enhanced by almost 45 % for most of the station while the RMSE decreased in the range from 3.4 mm to 0.5 mm. Moreover, corrected products could capture the true rainfall with a 30 % higher accuracy than the raw products as shown by POD and CSI calculations. This study contributes to the field of hydrology and climate science by improving the accuracy of satellite-derived precipitation estimates in mountainous regions and sets a precedent for the application of advanced machine learning techniques in environmental science. The implications of this research are far-reaching, offering valuable insights for water resource management, agricultural planning, and disaster mitigation efforts in the UAE and similar regions worldwide.</div></div>","PeriodicalId":362,"journal":{"name":"Journal of Hydrology","volume":"653 ","pages":"Article 132801"},"PeriodicalIF":5.9,"publicationDate":"2025-02-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143286842","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-01DOI: 10.1016/j.jhydrol.2024.132468
Dario Pumo, Francesco Alongi, Carmelo Nasello, Leonardo V. Noto
Remote sensing techniques for river monitoring facilitate faster measurement campaigns compared to traditional methods, reduce risks to personnel and instruments, and allow measurements under critical flow conditions. An alpha coefficient (α) is commonly employed to convert surface velocities, obtained by contactless techniques, into depth-averaged velocities, which are used for the application of the velocity-area method for assessing discharge. Some optical-based software programs use a constant α value, based on a theoretical “standard”. However, analyses of empirical vertical velocity profiles in real cases reveal that α can significantly deviate from this standard due to various factors (roughness, turbulence, etc.).
This study analyzes several ADCP (Acoustic Doppler Current Profiler)-based measurements in Sicily, Italy, to explore factors influencing flow velocity distribution and potential errors from using the standard α for discharge estimation via surface velocity-based methods. The results confirmed substantial variability in α, which is functionally related to some geometric factors characterizing the cross-section shape and the specific vertical where the velocity profile is computed. The generated dataset of empirical α values is also used to implement an Artificial Neural Network (ANN), offering a straightforward tool suitable for non-contact techniques. The ANN predicts α at any vertical of a measurement transect as a function of variables however necessary for discharge assessment by non-intrusive methods, leading to depth-averaged velocity estimates from surface velocities that are more accurate than those derived from conventional approaches, as demonstrated by four test cases.
{"title":"A simplified method for estimating the alpha coefficient in surface velocity based river discharge measurements","authors":"Dario Pumo, Francesco Alongi, Carmelo Nasello, Leonardo V. Noto","doi":"10.1016/j.jhydrol.2024.132468","DOIUrl":"10.1016/j.jhydrol.2024.132468","url":null,"abstract":"<div><div>Remote sensing techniques for river monitoring facilitate faster measurement campaigns compared to traditional methods, reduce risks to personnel and instruments, and allow measurements under critical flow conditions. An alpha coefficient (α) is commonly employed to convert surface velocities, obtained by contactless techniques, into depth-averaged velocities, which are used for the application of the velocity-area method for assessing discharge. Some optical-based software programs use a constant α value, based on a theoretical “standard”. However, analyses of empirical vertical velocity profiles in real cases reveal that α can significantly deviate from this standard due to various factors (roughness, turbulence, etc.).</div><div>This study analyzes several ADCP (Acoustic Doppler Current Profiler)-based measurements in Sicily, Italy, to explore factors influencing flow velocity distribution and potential errors from using the standard α for discharge estimation via surface velocity-based methods. The results confirmed substantial variability in α, which is functionally related to some geometric factors characterizing the cross-section shape and the specific vertical where the velocity profile is computed. The generated dataset of empirical α values is also used to implement an Artificial Neural Network (ANN), offering a straightforward tool suitable for non-contact techniques. The ANN predicts α at any vertical of a measurement transect as a function of variables however necessary for discharge assessment by non-intrusive methods, leading to depth-averaged velocity estimates from surface velocities that are more accurate than those derived from conventional approaches, as demonstrated by four test cases.</div></div>","PeriodicalId":362,"journal":{"name":"Journal of Hydrology","volume":"648 ","pages":"Article 132468"},"PeriodicalIF":5.9,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142790113","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号