Search Results (3,603)

Accurate estimation of crop evapotranspiration (ET) is essential for achieving efficient agricultural water use in the North China Plain. Although machine learning techniques have demonstrated considerable potential for ET simulation, a systematic evaluation of model-architecture suitability and hyperparameter optimization strategies specifically for summer soybean ET estimation in this region is still lacking. To address this gap, we systematically compared several machine learning architectures and their hyperparameter optimization schemes to develop a high-accuracy daily ET model for summer soybean in the North China Plain. Synchronous observations from a large-scale weighing lysimeter and an automatic weather station were first used to characterize the day-to-day dynamics of soybean ET and to identify the key driving variables. Four algorithms—support vector regression (SVR), Random Forest (RF), extreme gradient boosting (XGBoost), and a stacking ensemble—were then trained for ET simulation, while Particle Swarm Optimization (PSO), Genetic Algorithms (GAs), and Randomized Grid Search (RGS) were employed for hyperparameter tuning. Results show that solar radiation (R_S), maximum air temperature (T_max), and leaf area index (LAI) are the dominant drivers of ET. The Stacking-PSO-F3 combination, forced with Rs, T_max, LAI, maximum relative humidity (RH_max), and minimum relative humidity (RH_min), achieved the highest accuracy, yielding R² values of 0.948 on the test set and 0.900 in interannual validation, thereby demonstrating excellent precision, stability, and generalizability. The proposed model provides a robust technical tool for precision irrigation and regional water resource optimization. Full article

Understanding the mechanisms of hydrocarbon migration, accumulation, and alteration, particularly how evolution controls these processes, is critical for exploring lithologic hydrocarbons in reservoirs. In the complex tectonic settings of the continental margin of the stable North China Craton, there is a significant presence of small yet highly prolific hydrocarbon reservoirs. The processes of hydrocarbon migration and accumulation are complex and thus represent an important research focus in geology. This study, based on core, logging, and seismic data and integrating fluid inclusion analysis, quantitative fluorescence techniques, and geochemical experiments, combines the shale smear factor and paleotectonic reconstructions to clarify the hydrocarbon accumulation episodes, migration pathways, and factors controlling reservoir adjustments in the Yanwu area of the Tianhuan Depression in the Ordos Basin, China. The results reveal three types of NE-trending left-lateral strike–slip faults: linear, left-stepping, and right-stepping. Shale Smear Factor (SSF) analysis confirms that these faults exhibit segmented opening behaviors, with SSF > 1.7 identified as the threshold for fault openness. Multiparameter geochemical tracing based on terpanes and steranes shows that lateral migration along fault zones dominates the preferential migration pathways for hydrocarbons. Fluid inclusion thermometry revealed homogenization temperatures within the 100–110 °C and 80–90 °C intervals, while the oil inclusions exhibit blue or blue-and-white fluorescence, reflecting early hydrocarbon charging and late-stage secondary migration. Integrated analysis indicates that during the late Early Cretaceous (105–90 Ma), hydrocarbons were charged upward through open segments of linear strike–slip fault zones in the northern study area, experiencing lateral migration and accumulation along high-permeability sand bodies and unconformities in the shallow strata. Since the Late Cretaceous (65 Ma-present), the regional tectonic framework has evolved from a west–high, east–low to a west–low, east–high configuration, inducing secondary hydrocarbon migration and leading to the remigration or even destruction of early-formed oil reservoirs. This study systematically demonstrates that fault activity and tectonic evolution control the accumulation and distribution of hydrocarbons in the region. These findings provide theoretical insights for hydrocarbon exploration in regions with complex tectonic evolution within stable cratonic basins. Full article

The expansion of renewables in modern power systems and the coordinated development of upstream and downstream industrial chains are promoting a shift on the utility side from traditional settlement by energy toward operation driven by data and models. Industrial electricity consumption data exhibit pronounced multi-scale temporal structures and sectoral heterogeneity, which makes unified long-term load and generation forecasting while maintaining accuracy, interpretability, and scalability a challenge. From a modern system identification perspective, this paper proposes a System Identification in Adaptive Bayesian Framework (SIABF) for medium- and long-term industrial load forecasting based on daily freeze electricity time series. By combining daily aggregation of high-frequency data, frequency domain analysis, sparse identification, and long-term extrapolation, we first construct daily freeze series from 15 min measurements, and then we apply discrete Fourier transforms and a spectral complexity index to extract dominant periodic components and build an interpretable sinusoidal basis library. A sparse regression formulation with ${\ell }_1$ regularization is employed to select a compact set of key basis functions, yielding concise representations of sector and enterprise load profiles and naturally supporting multivariate and joint multi-sector modeling. Building on this structure, we implement a state-space-implicit physics-informed Bayesian forecasting model and evaluate it on real data from three representative sectors, namely, steel, photovoltaics, and chemical, using one year of 15 min measurements. Under a one-month-ahead evaluation using one year of 15 min measurements, the proposed framework attains a Mean Absolute Percentage Error (MAPE) of 4.5% for a representative PV-related customer case and achieves low single-digit MAPE for high-inertia sectors, often outperforming classical statistical models, sparse learning baselines, and deep learning architectures. These results should be interpreted as indicative given the limited time span and sample size, and broader multi-year, population-level validation is warranted. Full article

Tomato leaf diseases cause substantial yield and quality losses worldwide, yet reliable detection in real fields remains challenging. Two practical bottlenecks dominate current research: (i) limited data, including small samples for rare diseases, class imbalance, and noisy field images, and (ii) multimodal heterogeneity, where RGB images, textual symptom descriptions, spectral cues, and optional molecular assays provide complementary but hard-to-align evidence. This review summarizes recent advances in tomato leaf disease detection under these constraints. We first formalize the problem settings of limited and multimodal data and analyze their impacts on model generalization. We then survey representative solutions for limited data (transfer learning, data augmentation, few-/zero-shot learning, self-supervised learning, and knowledge distillation) and multimodal fusion (feature-, decision-, and hybrid-level strategies, with attention-based alignment). Typical model–dataset pairs are compared, with emphasis on cross-domain robustness and deployment cost. Finally, we outline open challenges—including weak generalization in complex field environments, limited interpretability of multimodal models, and the absence of unified multimodal benchmarks—and discuss future opportunities toward lightweight, edge-ready, and scalable multimodal systems for precision agriculture. Full article

Groundwater, a critical resource for regional water security and public health, faces escalating threats from heavy metal contamination—a pressing environmental challenge worldwide. This study focuses on the central-south Hunan region of China, a mineral-rich, densely populated area characterized predominantly by non-point-source pollution, aiming to systematically unravel the spatial patterns, source contributions, and associated health risks of heavy metals in local groundwater. Based on 717 spring and well water samples collected in 2024, we determined pH and seven heavy metals (As, Cd, Pb, Zn, Fe, Mn, and Tl). By integrating hydrogeological zoning, lithology, topography, and river networks, the study area was divided into 11 assessment units, clearly revealing the spatial heterogeneity of heavy metals. The results demonstrate that exceedances of Cd, Pb, and Zn were sporadic and point-source-influenced, whereas As, Fe, Mn, and Tl showed regional exceedance patterns (e.g., Mn exceeded the standard in 9.76% of samples), identifying them as priority control elements. The spatial distribution of heavy metals was governed the synergistic effects of lithology, water–rock interactions, and hydrological structure, showing a distinct “acidic in the northeast, alkaline in the southwest” pH gradient. Combined application of the APCS-MLR and PMF models resolved five principal pollution sources: an acid-reducing-environment-driven release source (contributing 76.1% of Fe and 58.3% of Pb); a geogenic–anthropogenic composite source (contributing 81.0% of Tl and 62.4% of Cd); a human-perturbation-triggered natural Mn release source (contributing 94.8% of Mn); an agricultural-activity-related input source (contributing 60.1% of Zn); and a primary geological source (contributing 89.9% of As). Monte Carlo simulation-based health risk assessment indicated that the average hazard index (HI) and total carcinogenic risk (TCR) for all heavy metals were below acceptable thresholds, suggesting generally manageable risk. However, As was the dominant contributor to both non-carcinogenic and carcinogenic risks, with its carcinogenic risk exceeding the threshold in up to 3.84% of the simulated adult exposures under extreme scenarios. Sensitivity analysis identified exposure duration (ED) as the most influential parameter governing risk outcomes. In conclusion, we recommend implementing spatially differentiated management strategies: prioritizing As control in red-bed and granite–metamorphic zones; enhancing Tl monitoring in the northern and northeastern granite-rich areas, particularly downstream of the Mishui River; and regulating land use in brick-factory-dense riparian zones to mitigate disturbance-induced Mn release—for instance, through the enforcement of setback requirements and targeted groundwater monitoring programs. This study provides a scientific foundation for the sustainable management and safety assurance of groundwater resources in regions with similar geological and anthropogenic settings. Full article

The archaeological terraces of Petra (southern Jordan) have long been recognized for their role in agriculture and flood mitigation. Despite the dominance of fine-grained sediments behind many terrace walls, these systems exhibit high infiltration capacity and remarkable resistance to erosion. This study investigates the hydrological behavior of terrace-trapped sediments through detailed soil texture, aggregate stability, salinity, and chemical analyses across eight representative sites in and around Petra. Grain-size distributions derived from dry and wet sieving, supplemented by laser diffraction, reveal that dry sieving substantially overestimates sand content due to aggregation of fine particles into unstable peds. Wet analyses demonstrate that many terrace soils are clay- or sandy-clay-dominated yet remain highly permeable. Chemical indicators (nitrate, phosphate, potassium, pH, and salinity) further suggest that terracing enhances downward water movement and salt leaching irrespective of clay content. The nature of the terrace settings and their sediment structure (especially the coarse-grained framework) exerts a stronger control on hydrological functioning than texture alone. The results have direct implications for understanding ancient land management in Petra and for informing sustainable terracing practices in modern arid and semi-arid landscapes, as they are effective both in harvesting water and reducing sediment mobilization. Full article

Parametric tiered-seating design can be framed as a constrained multi-objective optimization problem in which a low-dimensional decision vector is evaluated by a deterministic operator with sequential feasibility rejection and visibility constraints. This study introduces an oracle-preserving, learning-assisted screening workflow, where a multi-output multilayer perceptron (MLP) is used only to prioritize candidates for evaluation. Here, multi-output denotes a single network trained to predict the full objective vector jointly. Candidates are sampled within bounded decision ranges and evaluated by an operator that propagates section-coupled geometric state and enforces hard clearance thresholds through a Vertical Sightline System (VSS), i.e., a deterministic row-wise sightline/clearance evaluator that enforces hard clearance thresholds. The oracle-evaluated set is reduced to its mixed-direction Pareto-efficient subset and filtered by feature-space proximity to a fixed validation reference using nearest-neighbor distances in standardized 11-dimensional features, yielding a robustness-oriented pool. A compact shortlist is derived via TOPSIS (Technique for Order Preference by Similarity to an Ideal Solution; used here strictly as a post-Pareto decision-support ranking rule), and preference uncertainty is assessed by Monte Carlo weight sampling from a symmetric Dirichlet distribution. In an archived run under a fixed oracle budget, 1235 feasible designs are evaluated, producing 934 evaluated Pareto solutions; proximity filtering retains 187 robust candidates and TOPSIS reports a traceable top-30 shortlist. Stability is supported by concentrated top-k frequencies under weight perturbations and by audits under single-feature-drop ablations and tested rounding precisions. Overall, the workflow enables reproducible multi-objective screening and reporting for feasibility-dominated seating design. Full article

The collision dynamics of binary seawater droplets are pivotal in marine engineering applications, like spray desalination and engine cooling. While high-fidelity simulations can resolve these dynamics, they are computationally prohibitive for rapid design and analysis. This study introduces the first interpretable machine learning (ML) framework to predict and elucidate the collision outcomes of head-on binary seawater droplets. A high-fidelity numerical dataset, generated via Modified Coupled Level Set-VOF (M-CLSVOF) simulations across a broad Weber number (We) range, serves as the foundation for training multiple classifiers. Among the tested algorithms, the Random Forest model achieved superior performance with 96.2% accuracy. The model’s predictions precisely identified the critical Weber number for the transition from coalescence to reflexive separation at We ≈ 22.3 for seawater. Moving beyond black-box prediction, we employed SHapley Additive exPlanations (SHAP) to quantitatively deconstruct the model’s decision-making process. SHAP analysis confirmed the dominance of the Weber number (75% contribution) and revealed the context-dependent role of the Reynolds number (25% contribution) in modulating the collision outcome. Furthermore, a comparative analysis with freshwater droplets quantified a 6% elevation in the critical Weber number for seawater, attributed to salinity-induced modifications in fluid properties. Finally, a machine-learned regime map in the We-Ohnesorge space was constructed, delineating the coalescence and separation boundaries. This work establishes ML as a powerful, interpretable surrogate model that not only delivers rapid, accurate predictions but also extracts fundamental physical insights, offering a valuable paradigm for optimizing marine spray systems. Full article

Anise (Pimpinella anisum L.) is one of the most important annual herbs of the Apiaceae family, widely cultivated in southern Spain. Their seeds are highly valued for culinary uses and for producing quality essential oils widely used in food and beverage products, as well as for industry, medicinal, and cosmetics applications. This study investigates the seed yield and essential oil content within a set of 50 anise accessions from worldwide origin, as well as their composition by GC–MS and GC–FID analysis. Accessions showed significant differences in the agronomic parameters measured, including plant height (cm), seed yield (kg ha⁻¹), and the Harvest Index (%), with accessions PA_87 (Spain), PA_47 (Greece), and PA_21 (unknown origin) being the most performant. Essential oil (EO) content varied between 0.8% and 5.7% across different genotypes, resulting in EO production values ranging from 0.1 to 300 kg ha⁻¹. Trans-anethole was identified as the dominant terpene, comprising 84.4% to 94.4% of the content, followed by eugenol (1.4% to 5.5%) and α-muurolene (1.4% to 7.2%). PCA analysis identified five distinct groups and one outlier, influenced by minor terpenes. Indeed, there was a strong negative correlation between estragole and pseudoisoeugenyl 2-methylbutyrate. This study underscores the significance of minor terpenes, which play crucial roles in defining unique aniseed chemotypes, allowing for the selection of cultivars optimized for specific uses in food, cosmetics, and pharmaceuticals. Additionally, these findings emphasize the impact of cultivar genetics on agronomic traits and EO profiles, suggesting the need for further research to optimize plant growth and yield and EO quality. Full article

Coastal lakes are highly vulnerable transitional systems in which sedimentological processes and benthic ecological conditions jointly control contaminant accumulation and preservation, particularly in densely urbanized settings. A robust understanding of the physical and ecological characteristics of bottom sediments is therefore essential for the correct interpretation of contaminant distributions, including those of potentially toxic metals. In this study, an integrated sedimentological–ecological approach was applied to Lake Ganzirri, a Mediterranean shallow coastal lake located in northeastern Sicily (Italy), where recent investigations have identified localized heavy metal anomalies in surface sediments. Sediment texture, petrographic and mineralogical composition, malacofaunal assemblages, and lake-floor morpho-bathymetry were systematically analysed using grain-size statistics, faunistic determinations, GIS-based spatial mapping, and bivariate and multivariate statistical methods. The modern lake bottom is dominated by bioclastic quartzo-lithic sands with low fine-grained fractions and variable but locally high contents of calcareous skeletal remains, mainly derived from molluscs. Sediments are texturally heterogeneous, consisting predominantly of coarse-grained sands with lenses of very coarse sand, along with gravel and subordinate medium-grained sands. Both sedimentological features and malacofaunal death assemblages indicate deposition under open-lagoon conditions characterized by brackish waters and relatively high hydrodynamic energy. Spatial comparison between sedimentological–ecological parameters and previously published heavy metal distributions reveals no significant correlations with metal hotspots. The generally low metal concentrations, mostly below regulatory threshold values, are interpreted as being favoured by the high permeability and mobility of coarse sediments and by energetic hydrodynamic conditions limiting fine-particle accumulation. Overall, the integration of sedimentological and ecological data provides a robust framework for interpreting contaminant patterns and offers valuable insights for the environmental assessment and management of vulnerable coastal lake systems, as well as for the understanding of modern lagoonal sedimentary processes. Full article

This paper presents a PSCAD-based analysis of short-circuit faults and protection characteristics in a real distribution-level microgrid that integrates a 1 MWh battery energy storage system (BESS) with a 500 kW power conversion system (PCS) and a 500 kW photovoltaic (PV) plant connected to a 22.9 kV feeder. While previous studies often rely on simplified inverter models, this paper addresses the critical gap by integrating actual manufacturer-defined control parameters and cable impedances. This allows for a precise analysis of sub-millisecond transient behaviors, which is essential for developing robust protection schemes in inverter-dominated microgrids. The PSCAD model is first verified under grid-connected steady-state operation by examining PV output, BESS power, and grid voltage at the point of common coupling. Based on the validated model, DC pole-to-pole faults at the PV and ESS DC links and a three-phase short-circuit fault at the low-voltage bus are simulated to characterize the fault current behavior of the grid, BESS and PV converters. The DC fault studies confirm that current peaks are dominated by DC-link capacitor discharge and are strongly limited by converter controls, while the AC three-phase fault is mainly supplied by the upstream grid. As an initial application of the model, an instantaneous current change rate (ICCR) algorithm is implemented as a dedicated DC-side protection function. For a pole-to-pole fault, the ICCR index exceeds the 100 A/ms threshold and issues a trip command within 0.342 ms, demonstrating the feasibility of sub-millisecond DC fault detection in converter-dominated systems. Beyond this example, the validated PSCAD model and associated data set provide a practical platform for future research on advanced DC/AC protection techniques and protection coordination schemes in real BESS–PV microgrids. Full article

Length-based stock assessment methods are widely applied in data-limited fisheries, yet the effects of how length-frequency data are temporally grouped prior to analysis remain poorly examined. Temporal grouping is routinely used to increase sample size and approximate equilibrium conditions, but it may also alter the size structure presented to assessment models and bias inference. In this study, we evaluate how alternative temporal grouping schemes influence stock status inference within a single length-based framework, using the length-based spawning potential ratio (LBSPR) model as a diagnostic tool. Using a 30-year length-frequency dataset from a tropical purse seine fishery in the Northeast Atlantic as an illustrative case, we applied LBSPR under four practice-relevant temporal grouping schemes: full-period pooling, a broad regime-based scheme, decadal blocks, and five-year blocks. Life history parameters and model settings were held constant across schemes to isolate the effect of temporal grouping. A sensitivity analysis of biological parameters was conducted for the finest temporal scheme to contextualise robustness. Results show that temporal grouping alone can substantially alter the inferred status of the illustrative case. The fully pooled scheme produced an apparently favourable status signal, whereas finer temporal groupings revealed extended periods of inferred reproductive depletion, followed by a more recent recovery. Sensitivity analyses indicate that, while biological parameter uncertainty influences the magnitude of estimates, it does not overturn the dominant effect of temporal grouping on inferred status patterns. This study demonstrates that temporal grouping is not a neutral preprocessing step but a structural decision with the potential to conceal or reveal exploitation signals in length-based assessments. We argue that temporal grouping should be treated as an explicit sensitivity dimension in data-limited assessment workflows. By shifting attention from stock-specific outcomes to data-structuring choices, this work provides practical guidance for improving transparency and robustness in length-based stock status inference. Full article

To accurately investigate the loss characteristics of fuel cell vehicles (FCVs) under actual operating conditions and enhance their power performance and economic efficiency, this study establishes a numerical model of the proton exchange membrane fuel cell (PEMFC) stack based on real-world data from the second-generation Mirai. The stack model incorporates leakage current losses and imposes a limit on maximum current density. Besides, this study analyzes the effects of operating parameters (PEM water content, hydrogen partial pressure, current density, oxygen partial pressure, and operating temperature) on stack power output, efficiency, and eco-performance coefficient (ECOP). Furthermore, Non-Dominated Sequential Genetic Algorithm (NSGA-II) is employed to optimize the PEMFC stack performance, yielding the optimal operating parameter set for FCV operation. Further simulations are conducted on dynamic performance characteristics of the second-generation Mirai under two typical driving cycles, evaluating the power performance and economy of the FCV before and after optimization. Results demonstrate that the established PEMFC stack model accurately analyzes the output performance of an actual FCV when compared with real-world performance test data from the second-generation Mirai. Through optimization, output power increases by 7.4%, efficiency improves by 1.95%, and ECOP rises by 3.84%, providing guidance for enhancing vehicle power performance and improving overall vehicle economy. This study provides a practical framework for enhancing the power performance and overall energy sustainability of fuel cell vehicles, contributing to the advancement of sustainable transportation. Full article

Artificial intelligence (AI), particularly machine learning and deep learning architectures, has been widely applied to support thyroid cancer diagnosis, but existing evidence on its performance and limitations remains scattered across techniques, tasks, and data types. This systematic review synthesizes recent work on knowledge extraction from heterogeneous imaging and clinical data for thyroid cancer diagnosis and detection published between 2021 and 2025. We searched eight major databases, applied predefined inclusion and exclusion criteria, and assessed study quality using the Newcastle–Ottawa Scale. A total of 150 primary studies were included and analyzed with respect to AI techniques, diagnostic tasks, imaging and non-imaging modalities, model generalization, explainable AI, and recommended future directions. We found that deep learning, particularly convolutional neural networks, U-Net variants, and transformer-based models, dominated recent work, mainly for ultrasound-based benign–malignant classification, nodule detection, and segmentation, while classical machine learning, ensembles, and advanced paradigms remained important in specific structured-data settings. Ultrasound was the primary modality, complemented by cytology, histopathology, cross-sectional imaging, molecular data, and multimodal combinations. Key limitations included diagnostic ambiguity, small and imbalanced datasets, limited external validation, gaps in model generalization, and the use of largely non-interpretable black-box models with only partial use of explainable AI techniques. This review provides a structured, machine learning-oriented evidence map that highlights opportunities for more robust representation learning, workflow-ready automation, and trustworthy AI systems for thyroid oncology. Full article

To evaluate the epidemiology of herpetic keratitis over a 15-year period at a tertiary care center in Switzerland, focusing on the relative incidence of herpes simplex virus (HSV)-1, HSV-2, and varicella zoster virus (VZV), gender distribution, and co-infections, we conducted a retrospective single-center analysis of polymerase chain reaction (PCR) assays from corneal and conjunctival scrapings of suspected herpetic keratitis at a tertiary referral hospital. Patient demographics, viral spectra, and microbiological co-infections were assessed. Between 2010 and 2025, we identified 9954 PCR assays from 2892 patients, with 482 samples testing positive for herpesvirus. HSV-1 was the most frequent pathogen (328 of 3358, 9.8%), followed by VZV (143 of 3112, 4.6%), HSV-2 (9 of 3290, 0.27%), and CMV (2 of 194, 1.0%). Triplet testing (simultaneous HSV-1, HSV-2, and VZV-PCR) enabled direct comparisons of relative incidence rates. We found 2913 triplet testing results, with a relative distribution in positive results of 65.4% for HSV-1, 32.5% for VZV, and 2.1% for HSV-2. HSV-1 keratitis had a statistically significant higher incidence in men (58.9%, p = 0.0044), while no sex difference was detected for VZV (47.9%, p = 0.6683), HSV-2 (33.3%, p = 0.5078), or CMV (100%, p = 0.500). Bilateral infections were present in two patients, and co-infections were detected as follows: 8 cases of HSV-1/VZV co-detection, 3 cases of Acanthamoeba, and 15 of fungi. HSV-1 was the overwhelmingly dominant cause of herpetic keratitis at our institution, occurring more than twice as frequently as VZV and vastly outnumbering HSV-2. The statistically significant higher incidence in men in HSV-1 keratitis suggests possible biological or sociodemographic influences, whereas co-infections highlight the complexity of corneal pathology in a referral setting. These findings underscore the importance of multiplex PCR testing for accurate pathogen detection and provide insights into the epidemiologic landscape of herpetic keratitis. Full article