Pub Date : 2025-02-11DOI: 10.1016/j.envsoft.2025.106372
Leland Scantlebury , Vivek Bedekar , Matthew J. Tonkin , Marinko Karanovic , Thomas Harter
Subsurface hydraulic properties, critical in the development of groundwater models, are often inferred from aquifer tests and complemented by geologic information. In alluvial aquifers in particular, well and boring logs can provide a three-dimensional distribution of the presence of coarse-grained and fine-grained sediment (texture) as an important mapping of heterogeneity often correlated with hydraulic properties. Texture2Par was developed to incorporate texture data in the estimation of aquifer parameters for groundwater models. The software aggregates and interpolates texture data to a model grid, calculates hydraulic conductivity and storage parameters, and writes input files for the MODFLOW and IWFM simulation codes. Texture2Par includes options to represent a depth-dependent decrease in hydraulic conductivity, hydrostratigraphic units representing different depositional environments, and pilot points to represent relationships between texture and aquifer properties that may vary throughout groundwater flow systems. The paper presents the underlying methods and the application of Texture2Par using synthetic and real-world examples.
{"title":"Texture2Par: A texture-driven tool for estimating subsurface hydraulic properties","authors":"Leland Scantlebury , Vivek Bedekar , Matthew J. Tonkin , Marinko Karanovic , Thomas Harter","doi":"10.1016/j.envsoft.2025.106372","DOIUrl":"10.1016/j.envsoft.2025.106372","url":null,"abstract":"<div><div>Subsurface hydraulic properties, critical in the development of groundwater models, are often inferred from aquifer tests and complemented by geologic information. In alluvial aquifers in particular, well and boring logs can provide a three-dimensional distribution of the presence of coarse-grained and fine-grained sediment (texture) as an important mapping of heterogeneity often correlated with hydraulic properties. Texture2Par was developed to incorporate texture data in the estimation of aquifer parameters for groundwater models. The software aggregates and interpolates texture data to a model grid, calculates hydraulic conductivity and storage parameters, and writes input files for the MODFLOW and IWFM simulation codes. Texture2Par includes options to represent a depth-dependent decrease in hydraulic conductivity, hydrostratigraphic units representing different depositional environments, and pilot points to represent relationships between texture and aquifer properties that may vary throughout groundwater flow systems. The paper presents the underlying methods and the application of Texture2Par using synthetic and real-world examples.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"187 ","pages":"Article 106372"},"PeriodicalIF":4.8,"publicationDate":"2025-02-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143453055","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-10DOI: 10.1016/j.envsoft.2025.106361
Natalie Douglas, Tristan Quaife, Ross Bannister
The study presented here evaluates the ability of the 4DEnVar data assimilation technique to estimate the parameters from synthetically generated observations from a simple carbon model. The method is particularly attractive in its speed and ease of use, and its avoidance in construction of adjoint or tangent linear model code. Additionally, the assimilation analysis step can be performed independently of ensemble generation; there is no need to integrate the 4DEnVar code with that of the underlying model, assuming parameters are static in time. The 4DEnVar method is capable of closely estimating the model parameters with increased certainty given that the ensemble produces a sufficient number of trajectories exhibiting behaviour seen in the observations. We find that the root mean squared error between trajectories and observations is significantly reduced when compared with the prior — in one case a 96% and 99% reduction in the biomass and soil pools respectively.
{"title":"Exploring a hybrid ensemble–variational data assimilation technique (4DEnVar) with a simple ecosystem carbon model","authors":"Natalie Douglas, Tristan Quaife, Ross Bannister","doi":"10.1016/j.envsoft.2025.106361","DOIUrl":"10.1016/j.envsoft.2025.106361","url":null,"abstract":"<div><div>The study presented here evaluates the ability of the 4DEnVar data assimilation technique to estimate the parameters from synthetically generated observations from a simple carbon model. The method is particularly attractive in its speed and ease of use, and its avoidance in construction of adjoint or tangent linear model code. Additionally, the assimilation analysis step can be performed independently of ensemble generation; there is no need to integrate the 4DEnVar code with that of the underlying model, assuming parameters are static in time. The 4DEnVar method is capable of closely estimating the model parameters with increased certainty given that the ensemble produces a sufficient number of trajectories exhibiting behaviour seen in the observations. We find that the root mean squared error between trajectories and observations is significantly reduced when compared with the prior — in one case a 96% and 99% reduction in the biomass and soil pools respectively.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"186 ","pages":"Article 106361"},"PeriodicalIF":4.8,"publicationDate":"2025-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143388455","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-07DOI: 10.1016/j.envsoft.2025.106356
Tabea S. Sonnenschein , Zhendong Yuan , Jibran Khan , Jules Kerckhoffs , Roel C.H. Vermeulen , Simon Scheider
Scenario microsimulations like agent-based models can account for feedbacks and spatio-temporal and social heterogeneity when projecting future intervention impacts. Addressing air pollution exposure requires traffic scenario models (i.e. of car-free zones). Traditional air pollution models do not meet all requirements for traffic scenario microsimulation: isolating traffic emission, integrating relevant dispersion moderators, while computationally efficient, interoperable and valid. We propose a hybrid model of land use regression-based baseline concentrations and on-road emissions in conjunction with cellular automata-based off-road dispersion. The model efficiently assesses air pollution, while accounting for meteorological and morphological dispersion processes. We calibrate using genetic algorithms and externally validate the model based on mobile measurements and fixed-site routine monitoring data of NO2 concentrations across Amsterdam. Our model achieves an external validation R2 of 0.60 and 0.48 s computation time in a 50 m × 50 m raster. Further, we successfully projected the NO2 reduction of the first Covid-19 lockdown traffic scenario (R2 0.57).
{"title":"Hybrid cellular automata-based air pollution model for traffic scenario microsimulations","authors":"Tabea S. Sonnenschein , Zhendong Yuan , Jibran Khan , Jules Kerckhoffs , Roel C.H. Vermeulen , Simon Scheider","doi":"10.1016/j.envsoft.2025.106356","DOIUrl":"10.1016/j.envsoft.2025.106356","url":null,"abstract":"<div><div>Scenario microsimulations like agent-based models can account for feedbacks and spatio-temporal and social heterogeneity when projecting future intervention impacts. Addressing air pollution exposure requires traffic scenario models (<em>i.e</em>. of car-free zones). Traditional air pollution models do not meet all requirements for traffic scenario microsimulation: isolating traffic emission, integrating relevant dispersion moderators, while computationally efficient, interoperable and valid. We propose a hybrid model of land use regression-based baseline concentrations and on-road emissions in conjunction with cellular automata-based off-road dispersion. The model efficiently assesses air pollution, while accounting for meteorological and morphological dispersion processes. We calibrate using genetic algorithms and externally validate the model based on mobile measurements and fixed-site routine monitoring data of NO2 concentrations across Amsterdam. Our model achieves an external validation R2 of 0.60 and 0.48 s computation time in a 50 m × 50 m raster. Further, we successfully projected the NO2 reduction of the first Covid-19 lockdown traffic scenario (R2 0.57).</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"186 ","pages":"Article 106356"},"PeriodicalIF":4.8,"publicationDate":"2025-02-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143377303","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-07DOI: 10.1016/j.envsoft.2025.106367
Zizheng Guo , Haojie Wang , Jun He , Da Huang , Yixiang Song , Tengfei Wang , Yuanbo Liu , Joaquin V. Ferrer
Accurate landslide susceptibility assessments (LSA) are crucial for civil protection and land use planning. This study introduces PSLSA v2.0 as an open-source Python package that can conduct LSA automatically. It integrates six sophisticated machine learning algorithms (C5.0, SVM, LR, RF, MLP, XGBoost), and allows arbitrary combinations of influencing factors to generate landslide susceptibility index (LSI). We demonstrate how factor contribution and hyperparameter optimization as additional outputs can enhance the model interpretability. We apply PSLSA to a case study focused from Linzhi City in the Tibetan Plateau of China, that has undergone significant engineering modifications on its slopes. The results reveal that slope and aspect are the dominant factors in determining landslide susceptibility. All the six algorithms have an accuracy of over 80%. Although the distribution patterns of LSI vary, the C5.0 model is set apart with the best performance. PSLSA provides a powerful tool for stakeholders especially the non-geohazard professionals.
{"title":"PSLSA v2.0: An automatic Python package integrating machine learning models for regional landslide susceptibility assessment","authors":"Zizheng Guo , Haojie Wang , Jun He , Da Huang , Yixiang Song , Tengfei Wang , Yuanbo Liu , Joaquin V. Ferrer","doi":"10.1016/j.envsoft.2025.106367","DOIUrl":"10.1016/j.envsoft.2025.106367","url":null,"abstract":"<div><div>Accurate landslide susceptibility assessments (LSA) are crucial for civil protection and land use planning. This study introduces PSLSA v2.0 as an open-source Python package that can conduct LSA automatically. It integrates six sophisticated machine learning algorithms (C5.0, SVM, LR, RF, MLP, XGBoost), and allows arbitrary combinations of influencing factors to generate landslide susceptibility index (LSI). We demonstrate how factor contribution and hyperparameter optimization as additional outputs can enhance the model interpretability. We apply PSLSA to a case study focused from Linzhi City in the Tibetan Plateau of China, that has undergone significant engineering modifications on its slopes. The results reveal that slope and aspect are the dominant factors in determining landslide susceptibility. All the six algorithms have an accuracy of over 80%. Although the distribution patterns of LSI vary, the C5.0 model is set apart with the best performance. PSLSA provides a powerful tool for stakeholders especially the non-geohazard professionals.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"186 ","pages":"Article 106367"},"PeriodicalIF":4.8,"publicationDate":"2025-02-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143379430","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
IF 4.8 2区 环境科学与生态学Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS
Pub Date : 2025-02-06DOI: 10.1016/j.envsoft.2025.106366
T. Kang , S. Lee , H. An , M. Kim , I. Kimura
This study introduces Deb2D, an advanced predictive model that combines Eulerian flow dynamics with Lagrangian driftwood movement to accurately simulate debris flows. It enhances the existing Deb2D framework (An et al., 2019) by integrating a driftwood dynamics module rewritten in C++ (Kang et al., 2020) and a user-friendly Graphical User Interface developed with QtCreator for setup and visualization of simulations. This improvement enables precise two-way interactions between driftwood and debris flows, ensuring detailed visualization of their dynamics. When applied to the 2011 Mt. Umyeon debris flow in South Korea, the model demonstrated high accuracy in replicating observed phenomena. Future developments will focus on adapting this model into a QGIS plugin to broaden its applicability and user base.
{"title":"Development of an advanced numerical simulation program considering debris flow and driftwood behavior","authors":"T. Kang , S. Lee , H. An , M. Kim , I. Kimura","doi":"10.1016/j.envsoft.2025.106366","DOIUrl":"10.1016/j.envsoft.2025.106366","url":null,"abstract":"<div><div>This study introduces Deb2D, an advanced predictive model that combines Eulerian flow dynamics with Lagrangian driftwood movement to accurately simulate debris flows. It enhances the existing Deb2D framework (An et al., 2019) by integrating a driftwood dynamics module rewritten in C++ (Kang et al., 2020) and a user-friendly Graphical User Interface developed with QtCreator for setup and visualization of simulations. This improvement enables precise two-way interactions between driftwood and debris flows, ensuring detailed visualization of their dynamics. When applied to the 2011 Mt. Umyeon debris flow in South Korea, the model demonstrated high accuracy in replicating observed phenomena. Future developments will focus on adapting this model into a QGIS plugin to broaden its applicability and user base.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"186 ","pages":"Article 106366"},"PeriodicalIF":4.8,"publicationDate":"2025-02-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143395430","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-06DOI: 10.1016/j.envsoft.2025.106351
Qin Zhao , Jiajun Liu , Xinwen Yang , Hongda Qi , Jie Lian
With the growing interest in data-driven methods, Graph Neural Networks (GNNs) have demonstrated strong performance in forecasting as a deep learning architecture. However, GNN-based methods typically construct the graph based solely on the distance between stations, and few methods introduce geographical factors that significantly affect the spatial dispersion of , leading to performance bottlenecks. Additionally, these methods often fail to process the dynamic wind–field data comprehensively, resulting in inaccurate dispersion graph construction. These shortcomings greatly limit the interpretability of GNN models in forecasting air pollution. To address these issues, we propose a deep learning method that combines Graph Convolution Network (GCN) with Long Short-Term Memory (LSTM), leveraging geographical information within a dynamic graph. The model captures spatial dependencies between monitoring stations using a dynamic directional graph derived from the wind–field data and a static graph to represent inherent geographical relationships. The combination of GCN and LSTM enables the extraction of both spatial and temporal correlations. The results of experiments suggest that our proposed model, which offers great interpretability, outperforms state-of-the-art methods, especially in 24, 30, and 36 hours forecasts.
{"title":"Spatiotemporal PM2.5 forecasting via dynamic geographical Graph Neural Network","authors":"Qin Zhao , Jiajun Liu , Xinwen Yang , Hongda Qi , Jie Lian","doi":"10.1016/j.envsoft.2025.106351","DOIUrl":"10.1016/j.envsoft.2025.106351","url":null,"abstract":"<div><div>With the growing interest in data-driven methods, Graph Neural Networks (GNNs) have demonstrated strong performance in <span><math><msub><mrow><mi>PM</mi></mrow><mrow><mn>2</mn><mo>.</mo><mn>5</mn></mrow></msub></math></span> forecasting as a deep learning architecture. However, GNN-based methods typically construct the graph based solely on the distance between stations, and few methods introduce geographical factors that significantly affect the spatial dispersion of <span><math><msub><mrow><mi>PM</mi></mrow><mrow><mn>2</mn><mo>.</mo><mn>5</mn></mrow></msub></math></span>, leading to performance bottlenecks. Additionally, these methods often fail to process the dynamic wind–field data comprehensively, resulting in inaccurate <span><math><msub><mrow><mi>PM</mi></mrow><mrow><mn>2</mn><mo>.</mo><mn>5</mn></mrow></msub></math></span> dispersion graph construction. These shortcomings greatly limit the interpretability of GNN models in forecasting air pollution. To address these issues, we propose a deep learning method that combines Graph Convolution Network (GCN) with Long Short-Term Memory (LSTM), leveraging geographical information within a dynamic graph. The model captures spatial dependencies between <span><math><msub><mrow><mi>PM</mi></mrow><mrow><mn>2</mn><mo>.</mo><mn>5</mn></mrow></msub></math></span> monitoring stations using a dynamic directional graph derived from the wind–field data and a static graph to represent inherent geographical relationships. The combination of GCN and LSTM enables the extraction of both spatial and temporal correlations. The results of experiments suggest that our proposed model, which offers great interpretability, outperforms state-of-the-art methods, especially in 24, 30, and 36 hours forecasts.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"186 ","pages":"Article 106351"},"PeriodicalIF":4.8,"publicationDate":"2025-02-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143336705","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Landslides threaten lives and infrastructure, making accurate inventories crucial for risk management. This study combines expert methods with machine learning to automate and validate landslide detection and timing using Sentinel-2 satellite imagery. We developed ExMAD (Expert-based Multi-temporal AI Detector), an open-source methodological framework (https://github.com/NewGeoProjects/ExMAD) to integrate artificial intelligence with human expertise to detect occurrence timing of a targeted landslide. A U-Net neural network was chosen to effectively test ExMAD in landslide detection over Sentinel-2 worldwide multitemporal satellite imagery sequences, and the model was tested through five evaluations. ExMAD was able to effectively extract timing of target landslides on Sentinel-2 images and was able to correctly detect the presence/absence of landslide, proving the suitability of AI systems in landslide temporal mapping task.
This research proves the potential of hybrid AI-human approaches for landslide risk assessment, integrate human expertise with machine learning offers promising advancements for remote and field mapping of landslide. Furthermore, the ExMAD methodology adheres to the European Union's Artificial Intelligence Act, stressing human oversight in high-risk AI applications to enhance trust, control, and efficiency in landslide inventory creation and risk management.
{"title":"ExMAD (Expert-based Multitemporal AI Detector): An open-source methodological framework for remote and field landslide inventory","authors":"Michele Licata, Stefano Faga , Giandomenico Fubelli","doi":"10.1016/j.envsoft.2025.106363","DOIUrl":"10.1016/j.envsoft.2025.106363","url":null,"abstract":"<div><div>Landslides threaten lives and infrastructure, making accurate inventories crucial for risk management. This study combines expert methods with machine learning to automate and validate landslide detection and timing using Sentinel-2 satellite imagery. We developed ExMAD (Expert-based Multi-temporal AI Detector), an open-source methodological framework (<span><span>https://github.com/NewGeoProjects/ExMAD</span><svg><path></path></svg></span>) to integrate artificial intelligence with human expertise to detect occurrence timing of a targeted landslide. A U-Net neural network was chosen to effectively test ExMAD in landslide detection over Sentinel-2 worldwide multitemporal satellite imagery sequences, and the model was tested through five evaluations. ExMAD was able to effectively extract timing of target landslides on Sentinel-2 images and was able to correctly detect the presence/absence of landslide, proving the suitability of AI systems in landslide temporal mapping task.</div><div>This research proves the potential of hybrid AI-human approaches for landslide risk assessment, integrate human expertise with machine learning offers promising advancements for remote and field mapping of landslide. Furthermore, the ExMAD methodology adheres to the European Union's Artificial Intelligence Act, stressing human oversight in high-risk AI applications to enhance trust, control, and efficiency in landslide inventory creation and risk management.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"186 ","pages":"Article 106363"},"PeriodicalIF":4.8,"publicationDate":"2025-02-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143350438","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Improvements in public geospatial datasets provide opportunities for deriving urban drainage networks and simulation models of these networks (UDMs). We present SWMManywhere, which leverages such datasets for generating synthetic UDMs and creating a Storm Water Management Model for any urban area globally. SWMManywhere's modular and parameterised approach enables customisation to explore hydraulicly feasible network configurations. Key novelties of our workflow are in network topology derivation that accounts for combined effects of impervious area and pipe slope. We assess SWMManywhere by comparing pluvial flooding, drainage network outflows, and design with known networks. The results demonstrate high quality simulations are achievable with a synthetic approach even for large networks. Our sensitivity analysis shows that manholes locations, outfalls, and underlying street network are the most sensitive parameters. We find widespread sensitivity across all parameters without clearly defined values that they should take, thus, recommending an uncertainty driven approach to synthetic drainage network modelling.
{"title":"SWMManywhere: A workflow for generation and sensitivity analysis of synthetic urban drainage models, anywhere","authors":"Barnaby Dobson , Tijana Jovanovic , Diego Alonso-Álvarez , Taher Chegini","doi":"10.1016/j.envsoft.2025.106358","DOIUrl":"10.1016/j.envsoft.2025.106358","url":null,"abstract":"<div><div>Improvements in public geospatial datasets provide opportunities for deriving urban drainage networks and simulation models of these networks (UDMs). We present SWMManywhere, which leverages such datasets for generating synthetic UDMs and creating a Storm Water Management Model for any urban area globally. SWMManywhere's modular and parameterised approach enables customisation to explore hydraulicly feasible network configurations. Key novelties of our workflow are in network topology derivation that accounts for combined effects of impervious area and pipe slope. We assess SWMManywhere by comparing pluvial flooding, drainage network outflows, and design with known networks. The results demonstrate high quality simulations are achievable with a synthetic approach even for large networks. Our sensitivity analysis shows that manholes locations, outfalls, and underlying street network are the most sensitive parameters. We find widespread sensitivity across all parameters without clearly defined values that they should take, thus, recommending an uncertainty driven approach to synthetic drainage network modelling.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"186 ","pages":"Article 106358"},"PeriodicalIF":4.8,"publicationDate":"2025-02-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143379431","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","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.envsoft.2024.106301
Shaokun He , YiBo Wang , Dimitri Solomatine , Xiao Li
Long-term water resource management involving multipurpose coordination requires robust decision-making in water infrastructure cases to cope with various types of uncertainties. Traditional robust optimization methods generally do not explicitly propagate input or parametric uncertainties into estimates of the robustness of solutions, which limits their ability to address uncertainty comprehensively across solution spaces. In this study, we introduce an explicit robust decision-making framework that blends multiobjective search, probabilistic analysis of robustness, and diagnostic verification tools to identify robust optimal solutions to external uncertainty. The proposed framework is illustrated on four diverse robustness formulations, which capture a wide variety of stakeholder attitudes from highly risk-averse to risk-neutral, for the primary operating objectives (hydropower production, water diversion, and hydrological alteration degree) in China's Hanjiang cascade reservoir system. By analyzing the Pareto front propagated from inflow uncertainty, it is found that optimal robust policies with a significantly higher degree of hydrological alteration are preferred in most formulations to achieve relatively lower joint uncertainty of hydropower and water diversion. These policies also yield sufficiently stable model performance in the case of an out-of-sample streamflow set during diagnostic verification. Furthermore, a comparative analysis of four different formulations suggests that a composite normalized robustness indicator (NRI) developed in this study to integrate various robustness metrics can achieve an effective balance for all considered objectives. These findings highlight the benefits of explicit robust optimization for managing hydrological uncertainties in multipurpose cascade reservoirs.
{"title":"An explicit robust optimization framework for multipurpose cascade reservoir operation considering inflow uncertainty","authors":"Shaokun He , YiBo Wang , Dimitri Solomatine , Xiao Li","doi":"10.1016/j.envsoft.2024.106301","DOIUrl":"10.1016/j.envsoft.2024.106301","url":null,"abstract":"<div><div>Long-term water resource management involving multipurpose coordination requires robust decision-making in water infrastructure cases to cope with various types of uncertainties. Traditional robust optimization methods generally do not explicitly propagate input or parametric uncertainties into estimates of the robustness of solutions, which limits their ability to address uncertainty comprehensively across solution spaces. In this study, we introduce an explicit robust decision-making framework that blends multiobjective search, probabilistic analysis of robustness, and diagnostic verification tools to identify robust optimal solutions to external uncertainty. The proposed framework is illustrated on four diverse robustness formulations, which capture a wide variety of stakeholder attitudes from highly risk-averse to risk-neutral, for the primary operating objectives (hydropower production, water diversion, and hydrological alteration degree) in China's Hanjiang cascade reservoir system. By analyzing the Pareto front propagated from inflow uncertainty, it is found that optimal robust policies with a significantly higher degree of hydrological alteration are preferred in most formulations to achieve relatively lower joint uncertainty of hydropower and water diversion. These policies also yield sufficiently stable model performance in the case of an out-of-sample streamflow set during diagnostic verification. Furthermore, a comparative analysis of four different formulations suggests that a composite normalized robustness indicator (<em>NRI</em>) developed in this study to integrate various robustness metrics can achieve an effective balance for all considered objectives. These findings highlight the benefits of explicit robust optimization for managing hydrological uncertainties in multipurpose cascade reservoirs.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"185 ","pages":"Article 106301"},"PeriodicalIF":4.8,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142825336","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","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