Search Results (1,321)

This study compares and evaluates the performance of a statistical model, Multiple Linear Regression (MLR), and a deep learning model, Long Short-Term Memory (LSTM), for predicting monthly mean temperature in the Han River Basin, South Korea. Predictor variables were dynamically selected based on lagged correlation analysis between climate indices and temperature over the past 40 years, identifying the top ten variables with the highest correlations for lag times ranging from 1 to 18 months. The MLR model was developed through stepwise regression with cross-validation, while the LSTM model was constructed using an 18-month input sequence to capture temporal dependencies in the data. Model performance was evaluated using percent bias (PBIAS), Nash–Sutcliffe efficiency (NSE), Pearson’s correlation coefficient (r), and tercile-based probability metrics. Both models reproduced the seasonal variability of monthly temperature with high accuracy (NSE > 0.97, r > 0.98). The LSTM model showed slightly higher predictive skill in several periods but also exhibited larger prediction variance, reflecting the sensitivity of nonlinear architectures to variations in predictor–response relationships. In contrast, the MLR model demonstrated more stable predictive behavior with narrower uncertainty bounds, particularly under low signal-to-noise conditions, owing to its structural simplicity. These findings indicate that the two approaches are complementary; the LSTM model better captures nonlinear temporal dynamics, while the MLR model provides interpretability and robustness. Future work will explore advanced hybrid architectures such as CNN–LSTM and Transformer-based models, as well as multi-model ensemble methods, to further enhance the accuracy and reliability of medium-range temperature prediction. Full article

Satellite precipitation products serve as valuable global data sources for hydrological modeling, yet their applicability across different hydrological models remains insufficiently explored. The distributed physics-informed deep learning model (DPDL), as a representative of emerging differentiable, physics-based hydrological models, requires a systematic evaluation of the suitability of multi-source precipitation products within its modeling framework. This study focuses on the Xiangjiang River Basin in southern China, where both a DPDL model and a Soil and Water Assessment Tool (SWAT) model were constructed. In addition, two model training strategies were designed: S1 (fixed parameters) and S2 (product-specific recalibration). Multiple precipitation products were used to drive both hydrological models, and their streamflow simulation performance was evaluated under different training schemes to analyze the compatibility between precipitation products and hydrological modeling frameworks. The results show that: (1) In the Xiangjiang River Basin of southern China, GSMaP demonstrated the best overall performance with a Critical Success Index of 0.70 and a correlation coefficient (Corr) of 0.79; IMERG-F showed acceptable accuracy with a Corr of 0.75 but had a relatively high false alarm rate (FAR) of 0.32; while CMORPH exhibited the most significant systematic underestimation with a relative bias (RBIAS) of −8.48%. (2) The DPDL model more effectively captured watershed hydrological dynamics, achieving a validation period correlation coefficient of 0.82 and a Nash–Sutcliffe efficiency (NSE) of 0.79, outperforming the SWAT model. However, the DPDL model showed a higher RBIAS of +16.69% during the validation period, along with greater overestimation fluctuations during dry periods, revealing inherent limitations of differentiable hydrological models when training samples are limited. (3) The S2 strategy (product-specific recalibration) improved the streamflow simulation accuracy for most precipitation products, with the maximum increase in the NSE coefficient reaching 15.8%. (4) The hydrological utility of satellite products is jointly determined by model architecture and training strategy. For the DPDL model, IMERG-F demonstrated the best overall robustness, while GSMaP achieved the highest accuracy under the S2 strategy. This study aims to provide theoretical support for optimizing differentiable hydrological modeling and to offer new perspectives for evaluating the hydrological utility of satellite precipitation products. Full article

In this study, a Support Vector Regression (SVR) model was developed to predict the rate of penetration (ROP) during tunnel excavation, and its hyperparameters were optimized using Grid Search (GS), Random Search (RS), and Bayesian Optimization (BO). The results indicate that BO reached the optimal parameter set with only 30–50 evaluations, whereas GS and RS required approximately 1000 evaluations. In addition, BO achieved the highest predictive accuracy (R² = 0.9625) while reducing the computational time from 25.83 s (GS) to 17.31 s. Compared with the baseline SVM model, the optimized SVR demonstrated high accuracy (R² = 0.9610–0.9625), strong stability (NSE = 0.9194–0.9231), and low error levels (MAE = 0.0927–0.1099), clearly highlighting the critical role of hyperparameter optimization in improving model performance. To enhance interpretability, a feature importance analysis was conducted using four machine learning methods: Random Forest (RF), Bagged Trees (BT), Support Vector Machines (SVM), and the Generalized Additive Model (GAM). The relative contributions of BI, UCS, ALPHA, and DPW to ROP were evaluated, providing clearer insight into the model’s decision-making process and enabling more reliable engineering interpretation. Overall, integrating hyperparameter optimization with feature importance analysis significantly improves both predictive performance and model explainability. The proposed approach offers a robust, generalizable, and scientifically sound framework for TBM operations and geotechnical modeling applications. Full article

For an alpine inland river basin affected by climate change, the interaction between groundwater (GW) and surface water (SW) within the watershed plays a crucial role in water resource management. To explore the bidirectional dynamic coupling of surface water and groundwater, this work adopted the extensively employed SWAT–MODFLOW model. Results indicate that statistical parameters including R² (0.81 for calibration periods and 0.79 for validation), NSE (0.79 for calibration periods and 0.75 for validation), RMSE (0.59~1.25 m), and PBIAS (15.21%) demonstrate the dependability of the SWAT–MODFLOW model in evaluating groundwater–surface water exchange processes within alpine inland river basins. Long-term monitoring data show that groundwater levels exhibited an upward trend, rising from 2895.35 m in 2005 to 2906.75 m in 2022. Notably, since 2018, groundwater levels have entered a period of being consistently above the long-term average. In terms of spatial distribution, the groundwater level patterns in 2005, 2010, and 2015 remained relatively consistent, marked by a west-to-east decreasing gradient. However, by 2020, this spatial distribution pattern shifted, marked by an east-to-west decreasing gradient. Meanwhile, our results reveal a pattern of upstream surface water recharge, bidirectional fluctuation in the middle reaches, and downstream groundwater-dominated recharge during the period of 2000~2023. During the 2000–2009 period, groundwater in sub5 received recharge from surface water, with the exchange rate ranging from −4987.75 to −374.82 m³/d. Conversely, during 2010–2023, groundwater in sub5 discharged into surface water, with the exchange rate ranging from 1136.75 to 56,646.56 m³/d. Moreover, there is seasonal variability in the SW–GW interchange relationship. In spring and summer, surface water primarily replenishes groundwater, whereas in autumn and winter, groundwater primarily replenishes surface water. This study provides a foundational method for assessing groundwater–surface water interactions in alpine inland river basins, which will contribute to the evaluation and management of local water resources. Full article

Soil erosion in the purple soil region presents severe challenges with complex driving mechanisms. At the same time, evaluation and prediction of runoff and sediment dynamics are lacking for natural vegetation restoration in bare areas. The Mann–Kendall and Pettitt tests were employed to identify abrupt shift points in runoff and sediment dynamics, utilizing monitoring data from the Suining Soil and Water Conservation Experimental Station over the period from 1984 to 2018. Therefore, the research periods were divided into a baseline period (1984–1992) and an evaluation period (1993–2018). Subsequently, encompassing rainfall, runoff, sediment, topography, soil properties, and vegetation parameters, a Water Erosion Prediction Project (WEPP) model was established to quantify the reduction benefits of runoff and sediment during the period of forest restoration. We found that the calibrated WEPP model demonstrated satisfactory performance based on Nash–Sutcliffe efficiency coefficients (NSE > 0.5) and determination coefficients (R² > 0.5) for runoff and sediment simulations. The WEPP model and double-mass curve analysis method revealed that forest restoration reduced runoff and sediment by more than 80%. It is recommended to implement artificial vegetation restoration before reaching the threshold for natural vegetation restoration to achieve soil and water conservation goals. Full article

Storage-oriented reservoir schemes are effective for large-scale hydrological modeling, yet two important limitations remain. First, although some reservoirs seasonally adjust flood storage capacity (FSC), no global study has examined whether constant or seasonally varying FSC performs better. Second, these schemes rely on empirical operational-zone parameterization, but its impact on simulation accuracy has never been systematically assessed. This study presents an open-source Python module integrating three leading storage-oriented schemes (S25, Z17, H22) with both constant and seasonally varying FSC options. Evaluated using daily observations from 289 global reservoirs via Nash-Sutcliffe Efficiency (NSE), constant FSC significantly outperforms seasonal variation, increasing median outflow NSE by 0.18–0.47 and reducing storage error magnitude by 38–61%, and is selected as optimal for 84% of reservoirs. Sensitivity analysis across eight alternative zoning schemes shows that, under constant FSC, outflow remains stable, whereas seasonal FSC sharply increases sensitivity. Storage simulation is more sensitive overall, yet constant FSC consistently yields the smallest errors. This work provides the first global comparison of FSC strategies and the first systematic assessment of operational zone parameter uncertainty. It strongly recommends constant FSC with H22 or S25 as the default for large-scale modeling. The released module offers a flexible, reproducible platform for the community. Future extensions may incorporate demand-driven rules and hybrid calibration to further improve performance in data-rich regions. Full article

Background: Periodontitis has been associated with systemic diseases such as cerebrovascular events. Emerging research highlights the potential role of the microbiome in intracranial aneurysm formation and rupture. Aims: We aimed to explore the associations among periodontal pathogens and the outcomes in patients with aneurysmal subarachnoid hemorrhage (aSAH). Materials and Methods: A total of 43 aSAH patients were enrolled. Clinical probing depth measurement and microbiological culture were performed for all participants. The markers of systemic immune response (IL-6, hsCRP) and brain injury (NSE, S100B) were measured between 24 and 48 h after admission. Development of delayed cerebral ischemia (DCI) as the primary and clinical outcome, based on modified Rankin Scale as secondary endpoints, comprised the chosen metrics. Results: A significant association was observed between patients with periodontal pocket depth PPD ≥ 5 mm (n = 28) and DCI, which developed in 19 patients (p = 0.007). In the subgroup of patients with PPD ≥ 5 mm significant associations were found between certain periodontal pathogens and DCI. Higher hsCRP (p = 0.05), IL-6 (p = 0.037) levels were observed in cases with periodontal pathogens, independent of the depth of the pocket, suggesting systemic inflammation. Conclusions: Elevated hsCRP and IL-6 levels, periodontal pocket depth ≥ 5 mm, and red-complex periodontal pathogens are associated with an increased risk of DCI after aSAH, suggesting a role for periodontal disease–related systemic inflammation in DCI risk stratification. Full article

Modeling the water cycle in the critical zone requires understanding interactions between the soil–vegetation–atmosphere compartments. Mechanistic modeling of soil water flow relies on the accurate determination of hydrodynamic parameters that control hydraulic conductivity and water retention curves. These parameters can be derived either using pedotransfer functions (PTFs), using soil properties obtained from field samples, or through inverse modeling, which allows the parameters to be adjusted to minimize differences between simulations and observations. While PTFs are widely used due to their simplicity, inverse modeling requires specific instrumentation and advanced numerical tools. This study, conducted at the Hydro-Geochemical Environmental Observatory (Strengbach forested catchment) in France, aims to determine the optimal hydrodynamic parameters for two contrasting forest plots, one dominated by spruce and the other by beech. The methodology integrates granulometric data across multiple soil layers to estimate soil parameters using PTFs (Rosetta). Water content and conductivity data were then corrected to account for soil stoniness, improving the KGE and NSE metrics. Finally, inverse parameter estimation based on water content measurements allowed for refinement of the evaluation of α, Ks, and n. This framework to estimate soil parameter was applied on different time periods to investigate the influence of the calibration chronicles on the estimated parameters. Results indicate that our methodology is efficient and that the optimal calibration period does not correspond to one with the most severe drought conditions; instead, a balanced time series including both wet and dry phases is preferable. Our findings also emphasize that KGE and NSE must be interpreted with caution, and that long simulation periods are essential for evaluating parameter robustness. Full article

The infiltration process is an essential element of the hydrological cycle and water management. To provide a consideration for selecting an infiltration model and setting parameter values in the Oum Zessar watershed, the effectiveness of four infiltration models—Horton, Philip, Kostiakov, and Green–Ampt—is systematically evaluated using infiltration rate data measured in several field locations. The constant infiltration rate (CIR) of several locations was assessed using the double-ring infiltrometer technique and juxtaposed with values derived from the models. The parametric equations of each model were calibrated using time-series infiltration data obtained from the experimental observations. Excel functions were used to simplify the intricate mathematical calculations of the parameters. The model’s accuracy was assessed using six statistical metrics: Root Mean Square Error (RMSE), Sum of Squared Errors (SSE), Standard Error (STD ERR), and bias, along with the highest values of Nash–Sutcliffe Efficiency (NSE) and correlation (CORR). The greatest values of NSE and CORR, along with the lowest values of RMSE, SSE, STD ERR, and bias, indicate the optimal model. Moreover, the Model Performance Index (MPI) was implemented to evaluate the effectiveness of the modules by providing a clear scoring system for the models. The obtained results indicated that Kostiakov model displays the optimal fitting values on all indicators and locations, and the Horton model showed the second-best fitting values in most of the indicators. Full article

Flash floods are among the most destructive hazards in northeastern Bangladesh, particularly in Sylhet district, where intense rainfall from the Meghalaya hills generates rapid inundation of low-lying areas. This study applies the Delft3D Flexible Mesh (FM) Suite to simulate flash flood inundation in the Surma River catchment and assess its potential for hazard mapping. Hydrological inputs were obtained from Bangladesh Water Development Board (BWDB) stations, combined with bathymetric surveys and a 10 m resolution DEM derived from remote sensing data. Model calibration and validation were performed using observed discharge and water level data at SW267 for the years 2019–2020 and verified for flood events in 2012, 2016, and 2017. The model achieved strong agreement with observed flows (R² > 0.9, NSE = 0.75–0.93), and the simulated inundation extent corresponded well with Sentinel-1A satellite-derived flood maps. Validation indicated that Delft3D FM can reasonably capture flash flood propagation and floodplain inundation patterns, including frequently affected areas, e.g., Sylhet Uposhohor. The results demonstrate the value of integrating hydrodynamic modeling with satellite-based validation for improved flood risk management. Findings highlight the potential of Delft3D FM to support early warning, urban planning, and disaster preparedness in flash flood-prone regions of Bangladesh. Full article

This study compares the performance of empirical and hybrid deep learning (DL) models in estimating daily potential evapotranspiration (PET) in the Nakdong River Basin (NRB), South Korea, with the FAO-56 Penman–Monteith (PM) method as a reference. Two empirical models, Priestley–Taylor (P-T) and Hargreaves–Samani (H-S), and two DL models, a standalone Long Short-Term Memory (LSTM) network and a hybrid Convolutional Neural Network Bidirectional LSTM with an attention mechanism, were trained on a meteorological dataset (1973–2024) across 13 meteorological stations. Four input combinations (C1, C2, C3, and C4) were tested to assess the model’s robustness under varying data availability conditions. The results indicate that empirical models performed poorly, with a basin-wide RMSE of 5.04–5.79 mm/day and negative NSE (−10.37 to −13.99), and are therefore poorly suited to NRB. In contrast, DL models achieved significant improvements in accuracy. The hybrid CNN-BiLSTM Attention Mechanism (C1) produced the highest performance, with R² = 0.820, RMSE = 0.672 mm/day, NSE = 0.820, and KGE = 0.880, which was better than the standalone LSTM (R² = 0.756; RMSE = 0.782 mm/day). The generalization of heterogeneous climates was also verified through spatial analysis, in which the NSE at the station level consistently exceeded 0.70. The hybrid DL model was found to be highly accurate in representing the temporal variability and seasonal patterns of PET and is therefore more suitable for operational hydrological modeling and water-resource planning in the NRB. Full article

In the Republic of Korea, fire outbreaks caused by electrical devices are one of the most frequent accidents, causing severe damage to human lives and infrastructure. The metropolitan police, The National Institute of Scientific Investigation, and the National Fire Research Institute conduct fire root-cause inspections to determine whether these fires are external or internal infrastructure fires. However, obtaining results is a complex process. In addition, the situation has been hampered by the lack of sufficient digital forensics and relevant programs. Apart from electrical devices, multi-sockets are among the main fire instigators. In this study, we aim to verify the feasibility of utilizing YOLO-based deep-learning object detection models for fire-cause inspection systems for multi-sockets. Particularly, we have created a novel image dataset of multi-socket fire causes with 3300 images categorized into the three classes of socket, both burnt-in and burnt-out. This data was used to train various models, including YOLOv4-csp, YOLOv5n, YOLOR-csp, YOLOv6, and YOLOv7-Tiny. In addition, we have proposed an improved YOLOv5n-SE by adding a squeeze-and-excitation network (SE) into the backbone of the conventional YOLOv5 network and deploying it into a two-stage detector framework with a first stage of socket detection and a second stage of burnt-in/burnt-out classification. From the experiment, the performance of these models was evaluated, revealing that our work outperforms other models, with an accuracy of 91.3% mAP@0.5. Also, the improved YOLOv5-SE model was deployed in a web browser application. Full article

Aging profoundly modifies neuronal responses to ischemia. We aimed to define age-dependent features of neuronal metabolism and cell death after ischemic stroke by assessing NeuN, NSE, and Caspase-3 in human cortical neurons and by comparing transcriptional activity within PI3K/Akt/mTOR and PI3K/Akt/FOXO3a pathways across age groups. The aim of this study was to determine age-dependent features of neuronal metabolism and cellular degradation in ischemic stroke based on immunohistochemical assessment of NeuN, NSE, and Caspase-3 markers in human cerebral cortex neurons, as well as to conduct a comparative analysis of gene expression in the PI3K/Akt/mTOR and PI3K/Akt/FOXO3a signaling pathways involved in the regulation of neuronal survival and apoptosis. For the investigation, frontal cortex autopsies from patients with ischemic stroke (n = 154; “young”, “middle” and “elderly”; death ≤7 days post-onset) were examined. Histology (hematoxylin–eosin) and Nissl staining were used for morphology and neuron counts. Multiplex immunofluorescence (NeuN, NSE, Caspase-3) quantified metabolically active and apoptotic neurons, and the percentage of Caspase-3⁺ among NeuN⁺ cells was calculated. qRT-PCR measured PIK3CA, AKT2, MTOR, and FOXO3A expression in the infarct border zone. Based on our results, neuronal density and NeuN/NSE expression declined with aging, and the fraction of Caspase-3⁺ among NeuN⁺ neurons in the penumbra rose (young 42%, middle 82%, elderly 89%). Morphologically “intact” penumbral neurons frequently lacked NeuN/NSE, revealing covert dysfunction. Young brains showed balanced activation of PI3K/Akt/mTOR and PI3K/Akt/FOXO3a, whereas elderly brains exhibited reduced Akt/mTOR activity with FOXO3A predominance, consistent with pro-apoptotic, inflammatory, and dysregulated autophagic signaling. Thus, aging markedly reduces neuronal metabolic activity and increases apoptotic death in the infarct border zone after ischemic stroke. In older patients, there is an almost complete loss of NeuN and NSE expression in penumbral neurons with robust activation of the caspase cascade, whereas younger patients retain a pool of metabolically active neurons. Age-dependent dysregulation of PI3K/Akt signaling—characterized by FOXO3a hyperactivation and mTOR suppression—further promotes apoptosis and dysregulated autophagy. These changes likely underlie the limited efficacy of standard neuroprotection in ischemic stroke and support the need for age-tailored neurotropic therapy aimed at enhancing pro-survival pathways within the infarct border zone. Full article

Accurate forecasting of groundwater level dynamics poses a critical challenge for sustainable water management in arid regions. However, the strong spatiotemporal heterogeneity inherent in groundwater systems and their complex interactions between natural processes and human activities often limit the effectiveness of conventional prediction methods. To address this, a hybrid CNN-LSTM deep learning model is constructed. This model is designed to extract multivariate coupled features and capture temporal dependencies from multi-variable time series data, while simultaneously simulating the nonlinear and delayed responses of aquifers to groundwater abstraction. Specifically, the convolutional neural network (CNN) component extracts the multivariate coupled features of hydro-meteorological driving factors, and the long short-term memory (LSTM) network component models the temporal dependencies in groundwater level fluctuations. This integrated architecture comprehensively represents the combined effects of natural recharge–discharge processes and anthropogenic pumping on the groundwater system. Utilizing monitoring data from 2021 to 2024, the model was trained and tested using a rolling time-series validation strategy. Its performance was benchmarked against traditional models, including the autoregressive integrated moving average (ARIMA) model, recurrent neural network (RNN), and standalone LSTM. The results show that the CNN-LSTM model delivers superior performance across diverse hydrogeological conditions: at the upstream well AJC-7, which is dominated by natural recharge and discharge, the Nash–Sutcliffe efficiency (NSE) coefficient reached 0.922; at the downstream well AJC-21, which is subject to intensive pumping, the model maintained a robust NSE of 0.787, significantly outperforming the benchmark models. Further sensitivity analysis reveals an asymmetric response of the model’s predictions to uncertainties in pumping data, highlighting the role of key hydrogeological processes such as delayed drainage from the vadose zone. This study not only confirms the strong applicability of the hybrid deep learning model for groundwater level prediction in data-scarce arid regions but also provides a novel analytical pathway and mechanistic insight into the nonlinear behavior of aquifer systems under significant human influence. Full article

Biochemical Oxygen Demand (BOD) is a key indicator of organic pollution and a proxy indicator reflecting organic loading that can indirectly influence eutrophication processes in aquatic systems. This study presents a spatially explicit, physically based modeling framework for simulating BOD dynamics in the urbanizing Chao Phraya and Tha Chin Rivers Basin in central Thailand. The framework integrates the Personal Computer Storm Water Management Model (PCSWMM) with GIS-based datasets to represent pollutant sources, hydraulic flow, and land use. The model was calibrated and validated using data from 36 monitoring stations (2021–2022), achieving strong performance: an NSE of 0.72 and an MAE of 0.35 mg/L for the Chao Phraya River, and an NSE of 0.88 and an MAE of 0.12 mg/L for the Tha Chin River. Scenario simulations for 2032 projected BOD concentrations exceeded 4 mg/L in several downstream segments under the baseline (no-intervention) scenario, indicating elevated organic pollution and potential oxygen depletion that may indirectly exacerbate eutrophication risk in the Upper Gulf of Thailand, particularly in tidal zones with low dilution and nutrient accumulation. Model projections suggest that effective mitigation would require a 20–30% reduction in BOD loads, achievable through enhanced wastewater treatment and stricter pollution controls. Although BOD reduction alone cannot eliminate eutrophication, it supports broader nutrient management efforts by improving baseline water quality conditions. The proposed model offers a robust tool for identifying pollution hotspots, evaluating management strategies, and informing integrated river basin policies under continued urban growth. Full article