Pub Date : 2026-03-01Epub Date: 2026-02-14DOI: 10.1016/j.envsoft.2026.106915
Mengfan Teng , Miaomiao Liang , Shuo Wang , Yu Ding
Fine particulate matter (PM2.5) poses serious threats to public health and the environment. Current satellite-based PM2.5 estimates often lack nighttime data, leading to significant temporal discontinuities. To overcome this, this study developed a novel graph-based spatiotemporal deep neural network (G-STDNN) that generates seamless hourly PM2.5 concentrations across China. We first produced continuous daytime and nighttime aerosol optical depth (AOD) by filling missing Himawari-8 AHI data with MERRA-2 AOD. This improved AOD, combined with ERA5 meteorology, TROPOMI NO2, nighttime light, and geographical data, served as model input. The G-STDNN effectively captures complex spatiotemporal patterns of air pollution. For 2019–2020, the model demonstrated high accuracy in sample-based (R2 = 0.942, RMSE = 10.81 μg/m3). Using the filled AOD significantly improved estimation performance (R2value increased from 0.74 to 0.85). Nighttime estimates remained robust (R2 ≈ 0.84). This study provides a continuous, high-accuracy hourly PM2.5 dataset essential for exposure assessment and air quality management in China.
{"title":"Seamless hourly PM2.5 mapping across China with a graph spatiotemporal deep neural network","authors":"Mengfan Teng , Miaomiao Liang , Shuo Wang , Yu Ding","doi":"10.1016/j.envsoft.2026.106915","DOIUrl":"10.1016/j.envsoft.2026.106915","url":null,"abstract":"<div><div>Fine particulate matter (PM<sub>2.5</sub>) poses serious threats to public health and the environment. Current satellite-based PM<sub>2.5</sub> estimates often lack nighttime data, leading to significant temporal discontinuities. To overcome this, this study developed a novel graph-based spatiotemporal deep neural network (G-STDNN) that generates seamless hourly PM<sub>2.5</sub> concentrations across China. We first produced continuous daytime and nighttime aerosol optical depth (AOD) by filling missing Himawari-8 AHI data with MERRA-2 AOD. This improved AOD, combined with ERA5 meteorology, TROPOMI NO<sub>2</sub>, nighttime light, and geographical data, served as model input. The G-STDNN effectively captures complex spatiotemporal patterns of air pollution. For 2019–2020, the model demonstrated high accuracy in sample-based (R<sup>2</sup> = 0.942, RMSE = 10.81 μg/m<sup>3</sup>). Using the filled AOD significantly improved estimation performance (R<sup>2</sup>value increased from 0.74 to 0.85). Nighttime estimates remained robust (R<sup>2</sup> ≈ 0.84). This study provides a continuous, high-accuracy hourly PM<sub>2.5</sub> dataset essential for exposure assessment and air quality management in China.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"198 ","pages":"Article 106915"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146209273","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 : 2026-03-01Epub Date: 2026-01-20DOI: 10.1016/j.envsoft.2026.106880
Shaun S.H. Kim , Russell S. Crosbie , Warrick Dawes , Jai Vaze , Bill Wang , Cherry Mateo , Rebekah May , Sudeep Nair , Jahangir Alam
Basin-scale water resources models lack key physical factors such as antecedent conditions, that strongly influence transmission losses under dry conditions. This study presents the development and evaluation of two river transmission loss models for integration into river systems models: dynamic maximum alluvium as river storage (DMAARS) and DMAARS coupled with river dead storage (DMAARSDS). The models were applied to environmental flow events during the 2018/2019 drought in the northern Murray-Darling Basin and compared with a benchmark piecewise linear loss model. They provided significantly improved performance in 8 out of 12 fit metrics and more realistic estimates of environmental flow metrics. Scenario testing revealed that model choice significantly influences predictions, especially of baseline conditions and ecological benefits, e.g., peak water height, flow extent. Analyses also showed strong potential for use in long-term water resource planning. To enable adoption, the new models have been integrated into eWater Source as a community plugin.
{"title":"Improved river transmission loss modelling for environmental flow releases during droughts","authors":"Shaun S.H. Kim , Russell S. Crosbie , Warrick Dawes , Jai Vaze , Bill Wang , Cherry Mateo , Rebekah May , Sudeep Nair , Jahangir Alam","doi":"10.1016/j.envsoft.2026.106880","DOIUrl":"10.1016/j.envsoft.2026.106880","url":null,"abstract":"<div><div>Basin-scale water resources models lack key physical factors such as antecedent conditions, that strongly influence transmission losses under dry conditions. This study presents the development and evaluation of two river transmission loss models for integration into river systems models: dynamic maximum alluvium as river storage (DMAARS) and DMAARS coupled with river dead storage (DMAARSDS). The models were applied to environmental flow events during the 2018/2019 drought in the northern Murray-Darling Basin and compared with a benchmark piecewise linear loss model. They provided significantly improved performance in 8 out of 12 fit metrics and more realistic estimates of environmental flow metrics. Scenario testing revealed that model choice significantly influences predictions, especially of baseline conditions and ecological benefits, e.g., peak water height, flow extent. Analyses also showed strong potential for use in long-term water resource planning. To enable adoption, the new models have been integrated into eWater Source as a community plugin.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"198 ","pages":"Article 106880"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146014636","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}
Effective spatial monitoring of carbon fluxes is crucial for implementing climate change mitigation and adaptation measures. This study develops an advanced machine learning (ML) pipeline to assess integral carbon fluxes at regional scales using Earth observation data and ground-based measurements. We aimed to address main limitations of spatial ML assessments associated with ignorance of environmental processes’ physical nature. We propose a training pipeline ensuring prediction robustness and model generalization, introducing influential features and ground truth data selection strategy. This results in a robust mapping tool with uncertainty estimations, supported by Shapley values-based feature importance analysis for interpretability and physical meaning. Our approach utilizes data from 168 FLUXNET stations, NASA POWER meteorological reanalysis, and MODIS satellite observations to train a CatBoost gradient boosting model. The model achieves of 0.76 predicting monthly NEE values with high spatial–temporal coherence, opening possibilities for comprehensive terrestrial ecosystem carbon dynamics assessments.
{"title":"Data-driven approach to robust spatio-temporal assessment of carbon fluxes using Earth observation and ground-based data","authors":"Artem Gorbarenko , Mikhail Gasanov , Elizaveta Gorbarenko , Polina Tregubova , Anna Petrovskaia , Usman Tasuev , Svetlana Illarionova , Dmitrii Shardrin , Evgeny Burnaev","doi":"10.1016/j.envsoft.2026.106881","DOIUrl":"10.1016/j.envsoft.2026.106881","url":null,"abstract":"<div><div>Effective spatial monitoring of carbon fluxes is crucial for implementing climate change mitigation and adaptation measures. This study develops an advanced machine learning (ML) pipeline to assess integral carbon fluxes at regional scales using Earth observation data and ground-based measurements. We aimed to address main limitations of spatial ML assessments associated with ignorance of environmental processes’ physical nature. We propose a training pipeline ensuring prediction robustness and model generalization, introducing influential features and ground truth data selection strategy. This results in a robust mapping tool with uncertainty estimations, supported by Shapley values-based feature importance analysis for interpretability and physical meaning. Our approach utilizes data from 168 FLUXNET stations, NASA POWER meteorological reanalysis, and MODIS satellite observations to train a CatBoost gradient boosting model. The model achieves <span><math><msup><mrow><mi>R</mi></mrow><mrow><mn>2</mn></mrow></msup></math></span> of 0.76 predicting monthly NEE values with high spatial–temporal coherence, opening possibilities for comprehensive terrestrial ecosystem carbon dynamics assessments.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"198 ","pages":"Article 106881"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146014637","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 : 2026-03-01Epub Date: 2026-01-14DOI: 10.1016/j.envsoft.2026.106872
Jiahao Zhou , Sen He , Shanjunxia Wu , Jia Zhang , Qiuhua Wang , Fei Wang
High-quality observational data capturing the complete wildfire lifecycle are essential for validating and enhancing prediction models, yet such integrated datasets remain scarce. This study presents a modelling framework based on multitemporal data acquired through UAV sensing, including high-precision LiDAR, photogrammetry, and synchronous environmental monitoring. A multi-UAV relay observation strategy was developed to continuously record sub-second wildfire propagation dynamics. We demonstrate the utility of this framework through benchmark modelling experiments in fuel mapping, fire spread prediction, and burn severity assessment. The high-resolution data provide a valuable and comprehensive basis for evaluating model behavior across temporal scales, particularly in capturing early fire progression and fire-atmosphere interactions. It also reveals limitations in current modelling approaches. This work offers a robust resource for advancing wildfire environmental modelling.
{"title":"THU-Wildfire: A multitemporal, multimodal observation dataset for wildfire behavior dynamics","authors":"Jiahao Zhou , Sen He , Shanjunxia Wu , Jia Zhang , Qiuhua Wang , Fei Wang","doi":"10.1016/j.envsoft.2026.106872","DOIUrl":"10.1016/j.envsoft.2026.106872","url":null,"abstract":"<div><div>High-quality observational data capturing the complete wildfire lifecycle are essential for validating and enhancing prediction models, yet such integrated datasets remain scarce. This study presents a modelling framework based on multitemporal data acquired through UAV sensing, including high-precision LiDAR, photogrammetry, and synchronous environmental monitoring. A multi-UAV relay observation strategy was developed to continuously record sub-second wildfire propagation dynamics. We demonstrate the utility of this framework through benchmark modelling experiments in fuel mapping, fire spread prediction, and burn severity assessment. The high-resolution data provide a valuable and comprehensive basis for evaluating model behavior across temporal scales, particularly in capturing early fire progression and fire-atmosphere interactions. It also reveals limitations in current modelling approaches. This work offers a robust resource for advancing wildfire environmental modelling.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"198 ","pages":"Article 106872"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145995201","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}
Spatiotemporal prediction of ecological events informs management and mitigates negative impacts. However, the process is complex, requiring extensive data preprocessing, model selection, and evaluation. We introduce an automated spatio-temporal prediction package (STPredict) in Python that receives the raw data in various acceptable formats from the user, performs the preprocessing, including data imputation, selects from the covariates, chooses systematically from several default or user-defined predictive models, evaluates the performance of the final model, and makes a future prediction. As a case study, we demonstrate its use in predicting Mountain Pine Beetle infestations in Cypress Hills Park, Canada. Researchers can use STPredict to apply diverse types of models, including user-defined models, for predicting the time and location of ecological events with minimal effort. This automation not only reduces human error, but also allows ecologists to spend more time on improving, rather than implementing, the existing models.
{"title":"STPredict: A Python package automating spatio-temporal predictions","authors":"Arash Mari Oriyad , Arezoo Haratian , Mahdi Naderi , Nasrin Rafiei , Maryam Meghdadi , Zeinab Maleki , Pouria Ramazi","doi":"10.1016/j.envsoft.2026.106901","DOIUrl":"10.1016/j.envsoft.2026.106901","url":null,"abstract":"<div><div>Spatiotemporal prediction of ecological events informs management and mitigates negative impacts. However, the process is complex, requiring extensive data preprocessing, model selection, and evaluation. We introduce an automated spatio-temporal prediction package (STPredict) in Python that receives the raw data in various acceptable formats from the user, performs the preprocessing, including data imputation, selects from the covariates, chooses systematically from several default or user-defined predictive models, evaluates the performance of the final model, and makes a future prediction. As a case study, we demonstrate its use in predicting Mountain Pine Beetle infestations in Cypress Hills Park, Canada. Researchers can use STPredict to apply diverse types of models, including user-defined models, for predicting the time and location of ecological events with minimal effort. This automation not only reduces human error, but also allows ecologists to spend more time on improving, rather than implementing, the existing models.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"198 ","pages":"Article 106901"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146134550","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 : 2026-03-01Epub Date: 2026-02-04DOI: 10.1016/j.envsoft.2026.106883
Mahsa Hajihosseinlou , Abbas Maghsoudi , Reza Ghezelbash
Selecting appropriate hyperparameters is essential for achieving stable and reliable results in machine learning–based mineral prospectivity mapping (MPM). In this study, the AdaBoost algorithm was used to predict Pb–Zn mineral potential. AdaBoost was chosen for its ability to integrate multiple weak learners and enhance the recognition of underrepresented mineralization patterns within imbalanced datasets. Nevertheless, its performance may decrease in the presence of spatial heterogeneity and noisy data. To mitigate these issues, a hybrid Bayesian-Hyperband optimization strategy was applied to tune both the AdaBoost and its base learners. Bayesian optimization explores the hyperparameter space using probabilistic modeling, whereas Hyperband improves efficiency by allocating resources to promising configurations. The optimized model, trained on geological, geochemical, tectonic, and remote sensing data, demonstrated high predictive stability and spatial consistency, supporting its applicability in complex mineral systems.
{"title":"Development of a hybrid Bayesian-Hyperband optimization procedure: GeoAI-driven hyperparameter tuning of AdaBoost for enhancing Mineral Prospectivity Mapping","authors":"Mahsa Hajihosseinlou , Abbas Maghsoudi , Reza Ghezelbash","doi":"10.1016/j.envsoft.2026.106883","DOIUrl":"10.1016/j.envsoft.2026.106883","url":null,"abstract":"<div><div>Selecting appropriate hyperparameters is essential for achieving stable and reliable results in machine learning–based mineral prospectivity mapping (MPM). In this study, the AdaBoost algorithm was used to predict Pb–Zn mineral potential. AdaBoost was chosen for its ability to integrate multiple weak learners and enhance the recognition of underrepresented mineralization patterns within imbalanced datasets. Nevertheless, its performance may decrease in the presence of spatial heterogeneity and noisy data. To mitigate these issues, a hybrid Bayesian-Hyperband optimization strategy was applied to tune both the AdaBoost and its base learners. Bayesian optimization explores the hyperparameter space using probabilistic modeling, whereas Hyperband improves efficiency by allocating resources to promising configurations. The optimized model, trained on geological, geochemical, tectonic, and remote sensing data, demonstrated high predictive stability and spatial consistency, supporting its applicability in complex mineral systems.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"198 ","pages":"Article 106883"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146134552","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 : 2026-03-01Epub Date: 2026-02-11DOI: 10.1016/j.envsoft.2026.106912
Samuel Seuru , Volker Grimm , Michael Barton , Liliana Perez , Navid Mahdizadeh Gharakhanlou , Raja Sengupta , Alejandro Miguel Dagnino
Machine Learning (ML) is increasingly applied across environmental, social and interdisciplinary sciences to analyze complex systems and inform decision-making. Yet, this rapid growth has exposed significant gaps in methodological consistency, documentation, and reproducibility. The lack of standardized frameworks often leads to fragmented workflows and difficulties in interpreting or reproducing results across disciplines. To address these challenges, we introduce the ODE (Overview, Data, and Execution) protocol: a structured, accessible framework to support transparent documentation of ML workflows. Inspired by established standards such as ODD (Overview, Design concepts, Details), ODD + D (adding human Decision-making) for agent-based modeling, ODMAP (Overview, Data, Model, Assessment, Prediction) for species distribution models and FAIR (Findable, Accessible, Interoperable and Reusable) principles, ODE's novelty is to translate ML workflows into a standardized reporting format, specifying what must be described for transparency, reuse, and reproducibility. In practice, ODE is a reporting checklist, typically provided as supplementary material, supporting authors and reviewers.
{"title":"The ODE (Overview, Data, and Execution) protocol for a standardized use of machine learning in environmental, social and related interdisciplinary sciences","authors":"Samuel Seuru , Volker Grimm , Michael Barton , Liliana Perez , Navid Mahdizadeh Gharakhanlou , Raja Sengupta , Alejandro Miguel Dagnino","doi":"10.1016/j.envsoft.2026.106912","DOIUrl":"10.1016/j.envsoft.2026.106912","url":null,"abstract":"<div><div>Machine Learning (ML) is increasingly applied across environmental, social and interdisciplinary sciences to analyze complex systems and inform decision-making. Yet, this rapid growth has exposed significant gaps in methodological consistency, documentation, and reproducibility. The lack of standardized frameworks often leads to fragmented workflows and difficulties in interpreting or reproducing results across disciplines. To address these challenges, we introduce the ODE (Overview, Data, and Execution) protocol: a structured, accessible framework to support transparent documentation of ML workflows. Inspired by established standards such as ODD (Overview, Design concepts, Details), ODD + D (adding human Decision-making) for agent-based modeling, ODMAP (Overview, Data, Model, Assessment, Prediction) for species distribution models and FAIR (Findable, Accessible, Interoperable and Reusable) principles, ODE's novelty is to translate ML workflows into a standardized reporting format, specifying what must be described for transparency, reuse, and reproducibility. In practice, ODE is a reporting checklist, typically provided as supplementary material, supporting authors and reviewers.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"198 ","pages":"Article 106912"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146160925","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 : 2026-03-01Epub Date: 2026-01-13DOI: 10.1016/j.envsoft.2026.106877
Shaahin Nazarpour Tameh , Jennifer Drake , Anna Palla , Ilaria Gnecco
Bioretention cells (BRCs) are widely implemented to restore undeveloped hydrologic cycle; however, conventional BRCs need considerable surface area, limiting their applicability in densely populated areas. Compact BRCs like Filterra® have been designed to provide comparable hydrologic and pollutant removal effectiveness with a smaller footprint. The hydraulic characteristics of Filterra's engineered media were assessed through laboratory testing using KSAT and HYPROP devices and these results were integrated with field monitoring to implement a field-validated storm water management model (SWMM). Laboratory results showed a hydraulic conductivity of 1750 mm/h. The validated SWMM model replicated the outflow dynamics with satisfactory accuracy (KGE >0.35, R2 > 0.47), and the total suspended solids (TSS) removal was suitably predicted (R2 = 0.83). Results demonstrate that the field-validated SWMM model can be used to evaluate both hydrologic performance and pollutant TSS removal efficiency of compact BRCs, while noting its limitations in representing complex TSS dynamics.
{"title":"Compact bioretention cell for urban stormwater management: Assessment of hydrologic, hydraulic, and water quality performance via laboratory and SWMM modelling","authors":"Shaahin Nazarpour Tameh , Jennifer Drake , Anna Palla , Ilaria Gnecco","doi":"10.1016/j.envsoft.2026.106877","DOIUrl":"10.1016/j.envsoft.2026.106877","url":null,"abstract":"<div><div>Bioretention cells (BRCs) are widely implemented to restore undeveloped hydrologic cycle; however, conventional BRCs need considerable surface area, limiting their applicability in densely populated areas. Compact BRCs like Filterra® have been designed to provide comparable hydrologic and pollutant removal effectiveness with a smaller footprint. The hydraulic characteristics of Filterra's engineered media were assessed through laboratory testing using KSAT and HYPROP devices and these results were integrated with field monitoring to implement a field-validated storm water management model (SWMM). Laboratory results showed a hydraulic conductivity of 1750 mm/h. The validated SWMM model replicated the outflow dynamics with satisfactory accuracy (KGE >0.35, R<sup>2</sup> > 0.47), and the total suspended solids (TSS) removal was suitably predicted (R<sup>2</sup> = 0.83). Results demonstrate that the field-validated SWMM model can be used to evaluate both hydrologic performance and pollutant TSS removal efficiency of compact BRCs, while noting its limitations in representing complex TSS dynamics.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"198 ","pages":"Article 106877"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145962602","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 : 2026-03-01Epub Date: 2026-02-10DOI: 10.1016/j.envsoft.2026.106910
Hasitha Adikari , Christian O’Leary , Joe Harrington , Conor Lynch
Flash floods are among the most destructive hydrometeorological hazards, requiring forecasting approaches that balance accuracy, timeliness and scalability. Computer vision (CV) has become a key enabler, providing near-real-time information from unmanned aerial vehicles (UAVs), satellites and ground-based imagery. This structured narrative review synthesises advances across six domains relevant to forecasting workflows: flood extent mapping, debris and water-level detection, land use and land cover (LULC) classification, change detection, impact assessment, and image compression. Integration pathways between CV outputs and hydrological or hydraulic models are examined, revealing that such coupling remains limited in current literature. It also highlights the future areas of research in this domain. The systematic assessment shows that convolutional neural network (CNN)-based segmentation remains the most practical approach for extracting real-time information from image data, while transformers and lightweight models show promise for real-time use. Persistent challenges include the scarcity of UAV benchmarks, reproducibility gaps and weak operational integration.
{"title":"Computer vision in flash flood forecasting: A narrative review of applications, integration pathways, and future directions","authors":"Hasitha Adikari , Christian O’Leary , Joe Harrington , Conor Lynch","doi":"10.1016/j.envsoft.2026.106910","DOIUrl":"10.1016/j.envsoft.2026.106910","url":null,"abstract":"<div><div>Flash floods are among the most destructive hydrometeorological hazards, requiring forecasting approaches that balance accuracy, timeliness and scalability. Computer vision (CV) has become a key enabler, providing near-real-time information from unmanned aerial vehicles (UAVs), satellites and ground-based imagery. This structured narrative review synthesises advances across six domains relevant to forecasting workflows: flood extent mapping, debris and water-level detection, land use and land cover (LULC) classification, change detection, impact assessment, and image compression. Integration pathways between CV outputs and hydrological or hydraulic models are examined, revealing that such coupling remains limited in current literature. It also highlights the future areas of research in this domain. The systematic assessment shows that convolutional neural network (CNN)-based segmentation remains the most practical approach for extracting real-time information from image data, while transformers and lightweight models show promise for real-time use. Persistent challenges include the scarcity of UAV benchmarks, reproducibility gaps and weak operational integration.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"198 ","pages":"Article 106910"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146152966","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 : 2026-03-01Epub Date: 2026-02-10DOI: 10.1016/j.envsoft.2026.106911
Hongyuan Guo , Bingrui Chen , Rui Ma , Yihe Wang , Jianrong Zhu
High-resolution simulations of estuarine saltwater intrusion are computationally demanding and require efficient execution on heterogeneous computing platforms. In this study, the use of standard Fortran parallelization—DO CONCURRENT—to accelerate the unstructured quadrilateral grid finite-differencing estuarine and coastal ocean model (UFDECOM-i) within a unified codebase for both multicore CPUs and GPUs was investigated. Using the NVFORTRAN compiler, three versions were implemented: MC-UFDECOM-i on multicore CPUs, GPU-UFDECOM-i using automatic data migration, and GPUA-UFDECOM-i using lightweight OpenACC directives for explicit data management. The results show that DO CONCURRENT enables scalable shared-memory parallelism on CPUs, with speedups of up to 16.32 × , and provides functional portability to GPUs without code modification. However, optimal GPU performance requires explicit data management, with GPUA-UFDECOM-i reaching a maximum speedup of 21.48 × . These results demonstrate that DO CONCURRENT ensures portability and maintainability, whereas explicit data control remains essential for high GPU efficiency.
{"title":"Parallelization of the estuarine saltwater intrusion numerical forecast model UFDECOM-i using Fortran DO CONCURRENT","authors":"Hongyuan Guo , Bingrui Chen , Rui Ma , Yihe Wang , Jianrong Zhu","doi":"10.1016/j.envsoft.2026.106911","DOIUrl":"10.1016/j.envsoft.2026.106911","url":null,"abstract":"<div><div>High-resolution simulations of estuarine saltwater intrusion are computationally demanding and require efficient execution on heterogeneous computing platforms. In this study, the use of standard Fortran parallelization—DO CONCURRENT—to accelerate the unstructured quadrilateral grid finite-differencing estuarine and coastal ocean model (UFDECOM-i) within a unified codebase for both multicore CPUs and GPUs was investigated. Using the NVFORTRAN compiler, three versions were implemented: MC-UFDECOM-i on multicore CPUs, GPU-UFDECOM-i using automatic data migration, and GPUA-UFDECOM-i using lightweight OpenACC directives for explicit data management. The results show that DO CONCURRENT enables scalable shared-memory parallelism on CPUs, with speedups of up to 16.32 × , and provides functional portability to GPUs without code modification. However, optimal GPU performance requires explicit data management, with GPUA-UFDECOM-i reaching a maximum speedup of 21.48 × . These results demonstrate that DO CONCURRENT ensures portability and maintainability, whereas explicit data control remains essential for high GPU efficiency.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"198 ","pages":"Article 106911"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146152960","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