Pub Date : 2025-11-05DOI: 10.1016/j.envsoft.2025.106768
Changhong Hu , Songtao Ai , Shoukat Ali Shah , Xi Ding , Yi Cai , XinDe Chu , Hanxiao Yuan , Runchuan Ouyang , Meng Cui , Christo Pimpirev
Accurate forecast of sea ice concentration (SIC) is essential for understanding climate dynamics and forecasting future conditions. We propose the SpectraCast, a novel forecasting model combining Fast Fourier Transform (FFT), Least Squares (LS), and Autoregressive (AR) methods to forecast daily Arctic SIC. FFT extracts annual and semi-annual periodic components from SIC series, which are modelled using LS, while AR is applied to forecast residual components. Validation results indicate that SpectraCast achieves strong forecasting performance, with a mean absolute error below 6 % for one-year-ahead forecasts and a binary classification accuracy exceeding 93 % in identifying sea ice margins across all non-land Arctic grid cells. Additionally, we explore the impact of dominant frequency components on forecasting performance and extends the forecast horizon to 6 years. SpectraCast provides a robust, efficient method for long-term SIC forecasting, enhancing climate modelling and decision-making.
{"title":"A time series-based hybrid model for daily Arctic sea ice forecasting years in advance","authors":"Changhong Hu , Songtao Ai , Shoukat Ali Shah , Xi Ding , Yi Cai , XinDe Chu , Hanxiao Yuan , Runchuan Ouyang , Meng Cui , Christo Pimpirev","doi":"10.1016/j.envsoft.2025.106768","DOIUrl":"10.1016/j.envsoft.2025.106768","url":null,"abstract":"<div><div>Accurate forecast of sea ice concentration (SIC) is essential for understanding climate dynamics and forecasting future conditions. We propose the SpectraCast, a novel forecasting model combining Fast Fourier Transform (FFT), Least Squares (LS), and Autoregressive (AR) methods to forecast daily Arctic SIC. FFT extracts annual and semi-annual periodic components from SIC series, which are modelled using LS, while AR is applied to forecast residual components. Validation results indicate that SpectraCast achieves strong forecasting performance, with a mean absolute error below 6 % for one-year-ahead forecasts and a binary classification accuracy exceeding 93 % in identifying sea ice margins across all non-land Arctic grid cells. Additionally, we explore the impact of dominant frequency components on forecasting performance and extends the forecast horizon to 6 years. SpectraCast provides a robust, efficient method for long-term SIC forecasting, enhancing climate modelling and decision-making.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"196 ","pages":"Article 106768"},"PeriodicalIF":4.6,"publicationDate":"2025-11-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145447366","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 : 2025-11-04DOI: 10.1016/j.envsoft.2025.106770
Sajan Neupane, Jeffery S. Horsburgh, Razin Bin Issa, Sierra Young
Reliable, high-resolution streamflow data are essential for hydrologic research, flood forecasting, and water management. Camera-based monitoring provides a promising non-contact alternative to traditional sensors. However, deployment challenges exist in remote environments, including intermittent connectivity, limited storage, and the need for manual oversight. This paper presents a robust, cloud-integrated data acquisition workflow enabling autonomous operation of camera-based hydrologic stations. Built with low-cost Raspberry Pi computers and IP cameras, the system automates image capture, verifies file integrity during cloud upload, stores metadata locally, and deletes confirmed files to conserve storage. A serverless cloud monitoring component checks for expected uploads and triggers alerts on failure. The system includes failsafe mechanisms like scheduled reattempts, data integrity safeguards, and environment-variable-based configuration for flexibility. Deployed at two sites in Utah, it operated continuously for more than four months. This work offers a fault-tolerant, scalable solution for reliable imagery collection and transfer in large-scale environmental sensing networks.
{"title":"HydrocamCollect: A robust data acquisition and cloud data transfer workflow for camera-based hydrological monitoring","authors":"Sajan Neupane, Jeffery S. Horsburgh, Razin Bin Issa, Sierra Young","doi":"10.1016/j.envsoft.2025.106770","DOIUrl":"10.1016/j.envsoft.2025.106770","url":null,"abstract":"<div><div>Reliable, high-resolution streamflow data are essential for hydrologic research, flood forecasting, and water management. Camera-based monitoring provides a promising non-contact alternative to traditional sensors. However, deployment challenges exist in remote environments, including intermittent connectivity, limited storage, and the need for manual oversight. This paper presents a robust, cloud-integrated data acquisition workflow enabling autonomous operation of camera-based hydrologic stations. Built with low-cost Raspberry Pi computers and IP cameras, the system automates image capture, verifies file integrity during cloud upload, stores metadata locally, and deletes confirmed files to conserve storage. A serverless cloud monitoring component checks for expected uploads and triggers alerts on failure. The system includes failsafe mechanisms like scheduled reattempts, data integrity safeguards, and environment-variable-based configuration for flexibility. Deployed at two sites in Utah, it operated continuously for more than four months. This work offers a fault-tolerant, scalable solution for reliable imagery collection and transfer in large-scale environmental sensing networks.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"196 ","pages":"Article 106770"},"PeriodicalIF":4.6,"publicationDate":"2025-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145441552","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 : 2025-11-04DOI: 10.1016/j.envsoft.2025.106767
Zeinab Yavari, Amirreza Shahmiri, Mohammad Reza Nikoo
In situ water quality surveys are often costly and infrequent, leaving many reservoirs under-monitored. This study introduces a novel, integrated framework—featuring a no-code Google Earth Engine (GEE) toolkit and an automated Machine Learning (ML) pipeline—to facilitate accessible, localized monitoring. The methodology involves processing Landsat 8/9 and Sentinel-2 imagery via the GEE toolkit and synchronizing it with an initial dataset of 564 in situ Conductivity–Temperature–Depth (CTD) profiles from Oman's Wadi Dayqah Dam (WDD) reservoir, which resulted in 489 viable matchups for model training. The ML pipeline then automates feature selection and trains locally calibrated ensemble models to estimate Chlorophyll-a (Chl-a), turbidity, and Water Surface Temperature (WST). The calibrated models demonstrated strong predictive performance, achieving R2 values up to 0.84 for key parameters. This open-source framework empowers non-experts to conduct scalable, data-driven water quality assessments, offering an accessible solution for reservoir management in data-scarce regions.
{"title":"Accessible reservoir water quality monitoring: An integrated google earth engine and machine learning framework","authors":"Zeinab Yavari, Amirreza Shahmiri, Mohammad Reza Nikoo","doi":"10.1016/j.envsoft.2025.106767","DOIUrl":"10.1016/j.envsoft.2025.106767","url":null,"abstract":"<div><div>In situ water quality surveys are often costly and infrequent, leaving many reservoirs under-monitored. This study introduces a novel, integrated framework—featuring a no-code Google Earth Engine (GEE) toolkit and an automated Machine Learning (ML) pipeline—to facilitate accessible, localized monitoring. The methodology involves processing Landsat 8/9 and Sentinel-2 imagery via the GEE toolkit and synchronizing it with an initial dataset of 564 in situ Conductivity–Temperature–Depth (CTD) profiles from Oman's Wadi Dayqah Dam (WDD) reservoir, which resulted in 489 viable matchups for model training. The ML pipeline then automates feature selection and trains locally calibrated ensemble models to estimate Chlorophyll-a (Chl-a), turbidity, and Water Surface Temperature (WST). The calibrated models demonstrated strong predictive performance, achieving R<sup>2</sup> values up to 0.84 for key parameters. This open-source framework empowers non-experts to conduct scalable, data-driven water quality assessments, offering an accessible solution for reservoir management in data-scarce regions.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"196 ","pages":"Article 106767"},"PeriodicalIF":4.6,"publicationDate":"2025-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145441551","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 : 2025-11-03DOI: 10.1016/j.envsoft.2025.106764
Bao Liu , Jiaxin Li , Xianfu Jiang , Matthew Gibbs , Zirun Zhu , Xueqing Wang , Klaus Joehnk , Lei Gao
Accurate prediction of cyanobacterial blooms is essential to mitigate ecological and economic damage. Standalone models often fail to capture the nonlinear, multi-factor dynamics of cyanobacterial blooms, limiting prediction accuracy. To overcome this, we developed a hybrid deep-learning model integrating discrete wavelet transform (DWT), long short-term memory (LSTM) networks, autoregressive integrated moving average (ARIMA), and a particle swarm optimization-tuned artificial neural network (PSO-ANN) for residual compensation. DWT decomposes time-series into tendency and detail components, ARIMA forecasts linear trend of tendency, while single- and double-layer LSTMs predict nonlinear patterns of tendency and detail, respectively. Tested across five Murray River sites in Australia, our approach achieved root mean square errors of 647–3230 cells/ml, Nash-Sutcliffe efficiencies of 0.876–0.988, and Pearson correlation coefficients of 0.941–0.989, significantly outperforming random forest. Ablation analysis ranked model components by importance: DWT, high-frequency predictor (double-layer LSTM), low-frequency predictor (ARIMA combined with single-layer LSTM), and residual compensation (PSO-ANN).
{"title":"Enhancing riverine cyanobacterial bloom prediction: A hybrid deep learning approach combining wavelet decomposition, double-layer LSTM, ARIMA, and residual compensation","authors":"Bao Liu , Jiaxin Li , Xianfu Jiang , Matthew Gibbs , Zirun Zhu , Xueqing Wang , Klaus Joehnk , Lei Gao","doi":"10.1016/j.envsoft.2025.106764","DOIUrl":"10.1016/j.envsoft.2025.106764","url":null,"abstract":"<div><div>Accurate prediction of cyanobacterial blooms is essential to mitigate ecological and economic damage. Standalone models often fail to capture the nonlinear, multi-factor dynamics of cyanobacterial blooms, limiting prediction accuracy. To overcome this, we developed a hybrid deep-learning model integrating discrete wavelet transform (DWT), long short-term memory (LSTM) networks, autoregressive integrated moving average (ARIMA), and a particle swarm optimization-tuned artificial neural network (PSO-ANN) for residual compensation. DWT decomposes time-series into tendency and detail components, ARIMA forecasts linear trend of tendency, while single- and double-layer LSTMs predict nonlinear patterns of tendency and detail, respectively. Tested across five Murray River sites in Australia, our approach achieved root mean square errors of 647–3230 cells/ml, Nash-Sutcliffe efficiencies of 0.876–0.988, and Pearson correlation coefficients of 0.941–0.989, significantly outperforming random forest. Ablation analysis ranked model components by importance: DWT, high-frequency predictor (double-layer LSTM), low-frequency predictor (ARIMA combined with single-layer LSTM), and residual compensation (PSO-ANN).</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"196 ","pages":"Article 106764"},"PeriodicalIF":4.6,"publicationDate":"2025-11-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145435028","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 : 2025-10-30DOI: 10.1016/j.envsoft.2025.106757
Jun Zhu , Jianlin Wu , Yukun Guo , Jinbin Zhang , Zhihao Guo , Pei Dang , Weilian Li
As a vital application for geographic visualization, virtual geographic scenes can help the public enhance their risk awareness. However, existing research on virtual geographic scenes has excessively focused on visual effects, resulting in limited content diversity, inadequate expression of key information, and low cognitive efficiency for users. To address these issues, this article proposes a multi-level enhanced visualization method for virtual geographic scenes. First, a presentation-level enhancement method tailored for extended reality is developed. Second, an analysis-level enhancement method is designed by integrating collaborative semantic constraints and multiple visual variables. Third, an exploration-level enhancement method driven by interactive feedback is proposed to strengthen users' understanding of scenes and their cognitive exploration capabilities. Finally, a flood disaster was selected as a case study, and a prototype system was developed to support experimental analysis. The results indicate that the display-level enhancement significantly improves information transmission efficiency, the analysis-level enhancement markedly enhances cognitive accuracy and efficiency, and the exploration-level enhancement improves users’ decision-making support capabilities and cognitive engagement. These findings demonstrate that the proposed method enables users to capture key information in complex geographic scenes, enhances disaster risk perception, and provides scientific support for disaster management and emergency response.
{"title":"Multi-level enhanced visualization of virtual geographic scenes for improving risk perception: A case study of flood disaster scenes","authors":"Jun Zhu , Jianlin Wu , Yukun Guo , Jinbin Zhang , Zhihao Guo , Pei Dang , Weilian Li","doi":"10.1016/j.envsoft.2025.106757","DOIUrl":"10.1016/j.envsoft.2025.106757","url":null,"abstract":"<div><div>As a vital application for geographic visualization, virtual geographic scenes can help the public enhance their risk awareness. However, existing research on virtual geographic scenes has excessively focused on visual effects, resulting in limited content diversity, inadequate expression of key information, and low cognitive efficiency for users. To address these issues, this article proposes a multi-level enhanced visualization method for virtual geographic scenes. First, a presentation-level enhancement method tailored for extended reality is developed. Second, an analysis-level enhancement method is designed by integrating collaborative semantic constraints and multiple visual variables. Third, an exploration-level enhancement method driven by interactive feedback is proposed to strengthen users' understanding of scenes and their cognitive exploration capabilities. Finally, a flood disaster was selected as a case study, and a prototype system was developed to support experimental analysis. The results indicate that the display-level enhancement significantly improves information transmission efficiency, the analysis-level enhancement markedly enhances cognitive accuracy and efficiency, and the exploration-level enhancement improves users’ decision-making support capabilities and cognitive engagement. These findings demonstrate that the proposed method enables users to capture key information in complex geographic scenes, enhances disaster risk perception, and provides scientific support for disaster management and emergency response.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"196 ","pages":"Article 106757"},"PeriodicalIF":4.6,"publicationDate":"2025-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145382527","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 : 2025-10-30DOI: 10.1016/j.envsoft.2025.106760
Razi Sheikholeslami, Farshad Jahangiri
This study introduces a new Uncertainty-Informed Water Quality Index (UWQI) that directly incorporates data uncertainty into WQI calculations. The proposed UWQI integrates uncertainty-aware principal component analysis, Gaussian mixture models, and global sensitivity analysis to quantify and propagate uncertainty. As a case study, we applied the UWQI to 15 key WQ parameters from 27 monitoring stations across the Susquehanna River Basin in the northeastern US, from 2013 to 2021. Results show that neglecting uncertainty distorts WQ classifications considerably, leading to an 8 % increase in stations requiring treatment, a 19 % decline in ‘poor’ classifications, and a 26 % rise in ‘very poor’ stations. Our spatial analysis identified critical stations where data uncertainty notably influenced WQI assessments. Sensitivity analysis further revealed iron (Fe) data as the primary uncertainty source, followed by lead (Pb) and manganese (Mn). Our findings demonstrate that the proposed method can enhance WQI's transparency and credibility, minimizing misclassification and ineffective interventions.
{"title":"An uncertainty-informed water quality index: Incorporation of data uncertainty into water quality assessment","authors":"Razi Sheikholeslami, Farshad Jahangiri","doi":"10.1016/j.envsoft.2025.106760","DOIUrl":"10.1016/j.envsoft.2025.106760","url":null,"abstract":"<div><div>This study introduces a new Uncertainty-Informed Water Quality Index (UWQI) that directly incorporates data uncertainty into WQI calculations. The proposed UWQI integrates uncertainty-aware principal component analysis, Gaussian mixture models, and global sensitivity analysis to quantify and propagate uncertainty. As a case study, we applied the UWQI to 15 key WQ parameters from 27 monitoring stations across the Susquehanna River Basin in the northeastern US, from 2013 to 2021. Results show that neglecting uncertainty distorts WQ classifications considerably, leading to an 8 % increase in stations requiring treatment, a 19 % decline in ‘poor’ classifications, and a 26 % rise in ‘very poor’ stations. Our spatial analysis identified critical stations where data uncertainty notably influenced WQI assessments. Sensitivity analysis further revealed iron (Fe) data as the primary uncertainty source, followed by lead (Pb) and manganese (Mn). Our findings demonstrate that the proposed method can enhance WQI's transparency and credibility, minimizing misclassification and ineffective interventions.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"196 ","pages":"Article 106760"},"PeriodicalIF":4.6,"publicationDate":"2025-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145382528","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 : 2025-10-30DOI: 10.1016/j.envsoft.2025.106744
Jinbin Zhang, Hengchao Gu, Jun Zhu, Peiqi Que, Fanxu Huang, Yanhe Sun, Yuena Zhu, Weilian Li
The flood simulation in a 3D virtual geographic environment(VGE), which plays a key role in understanding and predicting floods, can provide a reliable basis for disaster prevention. However, current flood simulations employing VGE predominantly concentrate on large-scale scenes, such as rivers and urban areas. The existing research on flood simulation for fine-scale scenes exhibits a dearth of dynamic representation of flood spatiotemporal processes in 3D. Moreover, the failure of existing studies to adequately integrate data and models has resulted in difficulties in conducting fine-scale flood simulations within VGE. Therefore, this paper innovatively proposes a fine-scale flood simulation method driven by data and model coupling. First, a high-resolution surface grid was constructed by integrating multi-source data to provide a data foundation for fine-scale simulation. Second, a flood mechanism model based on improved cellular automata is developed for dynamically changing scenarios. Third, a fine-scale flood simulation method in the VGE was proposed by coupling the dynamic data layer with the mechanistic model layer. Finally, a prototype system was developed, which chose the urban underground station flooding as a typical case for experimental analysis. The results demonstrate that the proposed method achieves an 80 % simulation accuracy, which is highly consistent with field verification data and delivers excellent visualization effects. This approach innovatively extends flood simulation in VGE to fine-scale scenes such as underground stations, offering promising support for risk management in underground station construction during heavy rainfall scenarios.
{"title":"Coupling dynamic data and mechanistic models for fine-scale flood simulation in a 3D virtual geographic environment","authors":"Jinbin Zhang, Hengchao Gu, Jun Zhu, Peiqi Que, Fanxu Huang, Yanhe Sun, Yuena Zhu, Weilian Li","doi":"10.1016/j.envsoft.2025.106744","DOIUrl":"10.1016/j.envsoft.2025.106744","url":null,"abstract":"<div><div>The flood simulation in a 3D virtual geographic environment(VGE), which plays a key role in understanding and predicting floods, can provide a reliable basis for disaster prevention. However, current flood simulations employing VGE predominantly concentrate on large-scale scenes, such as rivers and urban areas. The existing research on flood simulation for fine-scale scenes exhibits a dearth of dynamic representation of flood spatiotemporal processes in 3D. Moreover, the failure of existing studies to adequately integrate data and models has resulted in difficulties in conducting fine-scale flood simulations within VGE. Therefore, this paper innovatively proposes a fine-scale flood simulation method driven by data and model coupling. First, a high-resolution surface grid was constructed by integrating multi-source data to provide a data foundation for fine-scale simulation. Second, a flood mechanism model based on improved cellular automata is developed for dynamically changing scenarios. Third, a fine-scale flood simulation method in the VGE was proposed by coupling the dynamic data layer with the mechanistic model layer. Finally, a prototype system was developed, which chose the urban underground station flooding as a typical case for experimental analysis. The results demonstrate that the proposed method achieves an 80 % simulation accuracy, which is highly consistent with field verification data and delivers excellent visualization effects. This approach innovatively extends flood simulation in VGE to fine-scale scenes such as underground stations, offering promising support for risk management in underground station construction during heavy rainfall scenarios.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"196 ","pages":"Article 106744"},"PeriodicalIF":4.6,"publicationDate":"2025-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145396810","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 : 2025-10-30DOI: 10.1016/j.envsoft.2025.106751
Dhiego da Silva Sales , David de Andrade Costa , Jader Lugon Junior , Ramiro Joaquim de Jesus Neves , Antonio José da Silva Neto
Accurate soil hydraulic parameters are essential for hydrological modeling, yet their spatial variability challenges parameterization. This study presents the MOHID SOIL TOOL (MST) to automate the integration of Brazilian Agricultural Research Corporation (EMBRAPA) soil texture data with Rosetta, an artificial neural network tool for estimating soil hydraulic parameters, enhancing hydrological simulations. The methodology involved programming automation routines to process soil data, ensuring compatibility with MOHID-Land adjusting soil hydraulic parameters to identify more realistic values that better represent local conditions. Developed in Python 3 with a Windows-compatible interface, MST automates the import, processing, and conversion of soil data. Testing in the Pedro do Rio watershed (Petrópolis, Brazil) demonstrated its efficiency in preparing soil input files for subsequent model calibration while reducing human errors. By optimizing workflow and ensuring precise data processing, MST advances hydrological research and supports sustainable water resource management, with flexibility for global raster-based soil datasets.
{"title":"Enhancing river flow predictions in MOHID-Land through integration of gridded soil data and hydraulic parameters using the MOHID SOIL TOOL","authors":"Dhiego da Silva Sales , David de Andrade Costa , Jader Lugon Junior , Ramiro Joaquim de Jesus Neves , Antonio José da Silva Neto","doi":"10.1016/j.envsoft.2025.106751","DOIUrl":"10.1016/j.envsoft.2025.106751","url":null,"abstract":"<div><div>Accurate soil hydraulic parameters are essential for hydrological modeling, yet their spatial variability challenges parameterization. This study presents the MOHID SOIL TOOL (MST) to automate the integration of Brazilian Agricultural Research Corporation (EMBRAPA) soil texture data with Rosetta, an artificial neural network tool for estimating soil hydraulic parameters, enhancing hydrological simulations. The methodology involved programming automation routines to process soil data, ensuring compatibility with MOHID-Land adjusting soil hydraulic parameters to identify more realistic values that better represent local conditions. Developed in Python 3 with a Windows-compatible interface, MST automates the import, processing, and conversion of soil data. Testing in the Pedro do Rio watershed (Petrópolis, Brazil) demonstrated its efficiency in preparing soil input files for subsequent model calibration while reducing human errors. By optimizing workflow and ensuring precise data processing, MST advances hydrological research and supports sustainable water resource management, with flexibility for global raster-based soil datasets.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"196 ","pages":"Article 106751"},"PeriodicalIF":4.6,"publicationDate":"2025-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145396811","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 : 2025-10-28DOI: 10.1016/j.envsoft.2025.106761
Alaa A. Masoud
Accurate hydrometric monitoring of reservoirs is critical for informed water resource management decisions. Lack of free software that can efficiently and rapidly, with effective reproducibility of the results, analyze big remote sensing data archives with no or little coding experience commonly hinder effective reservoir monitoring. Also, the paucity of independent and publicly available reservoir on-site storage data is another limit for accurate hydrometric evaluation of reservoirs. The study presents a Google Earth Engine (GEE) software for automated monitoring of surface water areas and levels, as well as storage curve estimation, using thresholding segmentation of timeseries radar (Sentinel-1 GRD) and optical (Landsat-8/9 and Sentinel-2 mosaics) images. A timeseries surface water area in the June 15, 2020–Jan. 25, 2025 period is estimated using regional thresholding and Otsu segmentation for the co-polarized (VV) and cross-polarized (VH) radar data, and thresholding for Sentinel-2 and Landsat-8/9 mosaics’ Normalized Difference Vegetation Index (NDVI) and Normalized Difference Water Index (NDWI). Water levels are estimated based on surface water area boundaries intersected with elevation from the ALOS PALSAR 12.5 m DEM near the reservoir dam. The maximum water level obtained is then utilized to estimate the elevation-storage curve. Water levels from various data and techniques are assessed as single-source, fused data of optical and radar, and the all-data composite to reference G-REALM reservoir altimeter data. The software has been validated for the Grand Ethiopian Reconnaissance Dam (GERD) reservoir over the course of five filling phases. A complete near real time water levels chart is produced at a shortened 6.4-day revisit time by a multi-sensor water levels fusion. Given the simplicity of the suggested strategy and the free availability of data, the study demonstrates that cloud-based remote sensing techniques can be used to fully automate large-scale hydrometric monitoring of water reservoirs.
对水库进行准确的水文监测对于明智的水资源管理决策至关重要。缺乏能够高效、快速、有效地再现结果的免费软件来分析没有或很少编码经验的大型遥感数据档案，通常会阻碍有效的油藏监测。此外，缺乏独立和公开的水库现场储存数据是对水库进行准确水文评价的另一个限制。该研究提出了一个谷歌Earth Engine （GEE）软件，用于自动监测地表水面积和水位，以及存储曲线估计，使用时间序列雷达（Sentinel-1 GRD）和光学（Landsat-8/9和Sentinel-2马赛克）图像的阈值分割。2020年6月15日- 2020年1月15日的时间序列地表水面积。利用共极化（VV）和交叉极化（VH）雷达数据的区域阈值和Otsu分割，以及Sentinel-2和Landsat-8/9马赛克的归一化植被指数（NDVI）和归一化差水指数（NDWI）的阈值，估算了2025年至2025年期间的数据。水位是根据与ALOS PALSAR 12.5 m DEM在水库大坝附近的高程相交的地表水区域边界估算的。然后利用得到的最大水位来估计高程-库容曲线。来自各种数据和技术的水位被评估为单源、光学和雷达融合数据，以及参考G-REALM水库高度计数据的所有数据组合。该软件已经在大埃塞俄比亚侦察大坝（GERD）水库的五个填充阶段进行了验证。通过多传感器水位融合，在缩短的6.4天重访时间内生成完整的近实时水位图。考虑到建议策略的简单性和数据的免费可用性，该研究表明，基于云的遥感技术可以用于完全自动化大规模水文监测水库。
{"title":"GEE-HydroMonitor: A Google Earth Engine software for multi-sensor hydrometric monitoring of surface water reservoirs","authors":"Alaa A. Masoud","doi":"10.1016/j.envsoft.2025.106761","DOIUrl":"10.1016/j.envsoft.2025.106761","url":null,"abstract":"<div><div>Accurate hydrometric monitoring of reservoirs is critical for informed water resource management decisions. Lack of free software that can efficiently and rapidly, with effective reproducibility of the results, analyze big remote sensing data archives with no or little coding experience commonly hinder effective reservoir monitoring. Also, the paucity of independent and publicly available reservoir on-site storage data is another limit for accurate hydrometric evaluation of reservoirs. The study presents a Google Earth Engine (GEE) software for automated monitoring of surface water areas and levels, as well as storage curve estimation, using thresholding segmentation of timeseries radar (Sentinel-1 GRD) and optical (Landsat-8/9 and Sentinel-2 mosaics) images. A timeseries surface water area in the June 15, 2020–Jan. 25, 2025 period is estimated using regional thresholding and Otsu segmentation for the co-polarized (VV) and cross-polarized (VH) radar data, and thresholding for Sentinel-2 and Landsat-8/9 mosaics’ Normalized Difference Vegetation Index (NDVI) and Normalized Difference Water Index (NDWI). Water levels are estimated based on surface water area boundaries intersected with elevation from the ALOS PALSAR 12.5 m DEM near the reservoir dam. The maximum water level obtained is then utilized to estimate the elevation-storage curve. Water levels from various data and techniques are assessed as single-source, fused data of optical and radar, and the all-data composite to reference G-REALM reservoir altimeter data. The software has been validated for the Grand Ethiopian Reconnaissance Dam (GERD) reservoir over the course of five filling phases. A complete near real time water levels chart is produced at a shortened 6.4-day revisit time by a multi-sensor water levels fusion. Given the simplicity of the suggested strategy and the free availability of data, the study demonstrates that cloud-based remote sensing techniques can be used to fully automate large-scale hydrometric monitoring of water reservoirs.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"195 ","pages":"Article 106761"},"PeriodicalIF":4.6,"publicationDate":"2025-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145382529","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}
Considerable research has been conducted to assess soil moisture (SM) of the upper soil layer by earth observation systems. Previous efforts have contributed to monitoring vegetation growth, status, and yield, which depend on SM. A more recent remote sensing-based model, the OPtical TRapezoid Model (OPTRAM), has been developed and validated. Unlike earlier thermal-based models, the current relies only on the commonly available visible and shortwave infrared spectral bands. The current paper presents an implementation of the OPTRAM model in the R programming language for assessing and mapping soil water content in rangelands worldwide. The new rOPTRAM package includes several innovations compared to early applications of the model. These new features include automated acquisition and pre-processing of Sentinel-2 imagery, programmatic delineation of trapezoid edges, and multiple curves (e.g., linear, exponential, and polynomial) fitting options for those edges.
{"title":"rOPTRAM: An R package for satellite-derived soil moisture in rangelands using the OPTRAM model","authors":"Micha Silver , Dong Zhe , Ricardo Diaz-Delgado , Arnon Karnieli","doi":"10.1016/j.envsoft.2025.106689","DOIUrl":"10.1016/j.envsoft.2025.106689","url":null,"abstract":"<div><div>Considerable research has been conducted to assess soil moisture (SM) of the upper soil layer by earth observation systems. Previous efforts have contributed to monitoring vegetation growth, status, and yield, which depend on SM. A more recent remote sensing-based model, the OPtical TRapezoid Model (OPTRAM), has been developed and validated. Unlike earlier thermal-based models, the current relies only on the commonly available visible and shortwave infrared spectral bands. The current paper presents an implementation of the OPTRAM model in the <span>R</span> programming language for assessing and mapping soil water content in rangelands worldwide. The new <span>rOPTRAM</span> package includes several innovations compared to early applications of the model. These new features include automated acquisition and pre-processing of Sentinel-2 imagery, programmatic delineation of trapezoid edges, and multiple curves (e.g., linear, exponential, and polynomial) fitting options for those edges.</div></div>","PeriodicalId":310,"journal":{"name":"Environmental Modelling & Software","volume":"195 ","pages":"Article 106689"},"PeriodicalIF":4.6,"publicationDate":"2025-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145382530","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