Search Results (112)

(1) Background: Wetlands are ecologically significant ecosystems that support biodiversity and contribute to essential environmental functions such as water purification, carbon storage and flood regulation. However, these ecosystems face increasing pressures from land-use change and degradation, prompting the need for scalable and accurate classification methods to support conservation and policy efforts. In this research, our motivation was to test whether high-spatial-resolution PlanetScope imagery can be used with pixel-based machine learning to support the mapping and monitoring of wetlands at a national scale. (2) Methods: This study compared four machine learning classification models—Random Forest (RF), XGBoost (XGB), Histogram-Based Gradient Boosting (HGB) and a Multi-Layer Perceptron Classifier (MLPC)—to detect and map wetland areas across New Zealand. All models were trained using eight-band SuperDove satellite imagery from PlanetScope, with a spatial resolution of ~3 m, and ancillary geospatial datasets representing topography and soil drainage characteristics, each of which is available globally. (3) Results: All four machine learning models performed well in detecting wetlands from SuperDove imagery and environmental covariates, with varying strengths. The highest accuracy was achieved using all eight image bands alongside features created from supporting geospatial data. For binary wetland classification, the highest F1 scores were recorded by XGB (0.73) and RF/HGB (both 0.72) when including all covariates. MLPC also showed competitive performance (wetland F1 score of 0.71), despite its relatively lower spatial consistency. However, each model over-predicts total wetland area at a national level, an issue which was able to be reduced by increasing the classification probability threshold and spatial filtering. (4) Conclusions: The comparative analysis highlights the strengths and trade-offs of RF, XGB, HGB and MLPC models for wetland classification. While all four methods are viable, RF offers some key advantages, including ease of deployment and transferability, positioning it as a promising candidate for scalable, high-resolution wetland monitoring across diverse ecological settings. Further work is required for verification of small-scale wetlands (<~0.5 ha) and the addition of fine-spatial-scale covariates. Full article

Accurate crop row detection is crucial for determining weed pressure (weeds item per square meter). However, this task is complicated by the similarity between crops and weeds, the presence of missing plants within rows, and the varying growth stages of both. Our hypothesis was that in drone imagery captured at altitudes of 20–30 m—where individual plant details are not discernible—weed presence among crops can be statistically detected, allowing for the generation of a weed distribution map. This study proposes a computer vision detection method using images captured by unmanned aerial vehicles (UAVs) consisting of six main phases. The method was tested on 208 images. The algorithm performs well under normal conditions; however, when the weed density is too high, it fails to detect the row direction properly and begins processing misleading data. To investigate these cases, 120 artificial datasets were created with varying parameters, and the scenarios were analyzed. It was found that a rate variable—in-row concentration ratio (IRCR)—can be used to determine whether the result is valid (usable) or invalid (to be discarded). The F1 score is a metric combining precision and recall using a harmonic mean, where “1” indicates that precision and recall are equally weighted, i.e., β = 1 in the general Fβ formula. In the case of moderate weed infestation, where 678 crop plants and 600 weeds were present, the algorithm achieved an F1 score of 86.32% in plant classification, even with a 4% row disturbance level. Furthermore, IRCR also indicates the level of weed pressure in the area. The correlation between the ground truth weed-to-crop ratio and the weed/crop classification rate produced by the algorithm is 98–99%. As a result, the algorithm is capable of filtering out heavily infested areas that require full weed control and capable of generating weed density maps on other cases to support precision weed management. Full article

Mapping marine habitats is fundamental for biodiversity conservation and ecosystem-based management in oceanic regions under increasing anthropogenic and climatic pressures. In the context of global initiatives—such as marine protected area expansion and international agreements—habitat mapping has become mandatory for regional and global conservation policies. It provides spatial data to delineate essential habitats, support connectivity analyses, and assess pressures, enabling ecosystem-based marine spatial planning aligned with EU directives (2008/56/EC; 2014/89/EU). Beyond biodiversity, macrophytes, rhodolith beds, and coral reefs deliver key ecosystem services—carbon sequestration, coastal protection, nursery functions, and fisheries support—essential to local socioeconomies. This systematic review (PRISMA guidelines) examined 69 peer-reviewed studies across Central-Eastern Atlantic archipelagos (Macaronesia: the Azores, Madeira, the Canaries, and Cabo Verde) and the Mid-Atlantic Ridge. We identified knowledge gaps, methodological trends, and key challenges, emphasizing the integration of cartographic, ecological, and technological approaches. Although methodologies diversified over time, the lack of survey standardization, limited ground truthing, and heterogeneous datasets constrained the production of high-resolution bionomic maps. Regional disparities persist in technology access and habitat coverage. The Azores showed the highest species richness (393), dominated by acoustic mapping in corals. Madeira was most advanced in the remote mapping of rhodoliths; the Canaries focused on shallow macrophytes with direct mapping; and Cabo Verde remains underrepresented. Harmonized protocols and regional cooperation are needed to improve data interoperability and predictive modeling. Full article

The Western Hemisphere (WH) circulation pattern was discovered in recent years through Self-Organizing Maps (SOMs) clustering of the Northern Hemisphere 500 hPa geopotential height during winter. For example, the extremely cold wave that occurred in North America during 2013–14 is associated with WH circulation anomalies. We discussed the extremely cold weather conditions within the WH pattern during the winter season from 1979 to 2023. The variations of cold air in North America during the WH pattern have been demonstrated using the NCEP/NCAR reanalysis datasets. By defining WH events and North American extremely cold events, we have identified a connection between the two. In extremely cold events, linear winds are the key factor driving the temperature drop, as determined by calculating temperature advection. The ridge in the Gulf of Alaska serves as an early signal for this cold weather. The WH circulation anomaly triggers an anomalous ridge in the Gulf of Alaska region, leading to trough anomalies downstream over North America. This results in the southward movement of cold air from the polar regions, causing cooling in the mid-to-northern parts of North America. With the maintenance of the stationary wave in the North Pacific (NP), the anomalous trough over North America can be deepened, driving cold air into the continent. Influenced by the low pressure over Greenland and the storm track, the cold anomalies are concentrated in the central and northern parts of North America. This cold air situation persists for approximately two weeks. The high-level patterns of the WH pattern in both the 500 hPa height and the troposphere level have been identified using SOM. This cold weather is primarily a tropospheric phenomenon with limited correlation to stratospheric activities. Full article

This study was conducted in the Harran Plain within the framework of the Southeastern Anatolia Project (GAP) in Türkiye to evaluate the vulnerability of groundwater to contamination, with a special emphasis on the high salinity conditions attributed to agricultural and rural practices. The region is notably challenged by salinization resulting from intensive irrigation and insufficient drainage systems. The DRASTIC framework was used to assess groundwater contamination vulnerability. The DRASTIC framework parameters were numerically integrated using both the original DRASTIC framework and its modified version, serving as the basis for subsequent predictive analytics and soft computing model development. The primary aim was to determine the most effective predictive model for groundwater contamination vulnerability in salinity-affected areas. In this context, various models were implemented and evaluated, including artificial neural networks (ANNs) with varied hidden layer configurations, four different regression-based methods (MARS, TreeNet, GPS, and CART), and three classical regression analysis approaches. The modeling process utilized 24 adjusted vulnerability indices (AVIs) as target variables, with the dataset partitioned into 58.34% for training, 20.83% for validating, and 20.83% for testing. Model performance was rigorously assessed using various statistical indicators such as mean absolute error, root mean square error, and the Nash–Sutcliffe efficiency coefficient, in addition to evaluating the predictive AVIs through spatial mapping. The findings revealed that the ANNs and TreeNet models offered superior performance in accurately predicting groundwater contamination vulnerability, particularly by delineating the spatial distribution of risk in areas experiencing intensive agricultural pressure. Full article

Significant challenges persist in diagnosing non-ischemic cardiomyopathies (NICMs) owing to early morphological overlap and subtle functional changes. While cardiac magnetic resonance (CMR) offers gold-standard structural assessment, current morphology-based AI models frequently overlook key biomechanical dysfunctions like diastolic/systolic abnormalities. To address this, we propose a dual-path hybrid deep learning framework based on CNN-LSTM and MLP, integrating anatomical features from cine CMR with biomechanical markers derived from intraventricular pressure gradients (IVPGs), significantly enhancing NICM subtype classification by capturing subtle biomechanical dysfunctions overlooked by traditional morphological models. Our dual-path architecture combines a CNN-LSTM encoder for cine CMR analysis and an MLP encoder for IVPG time-series data, followed by feature fusion and dense classification layers. Trained on a multicenter dataset of 1196 patients and externally validated on 137 patients from a distinct institution, the model achieved a superior performance (internal AUC: 0.974; external AUC: 0.962), outperforming ResNet50, VGG16, and radiomics-based SVM. Ablation studies confirmed IVPGs’ significant contribution, while gradient saliency and gradient-weighted class activation mapping (Grad-CAM) visualizations proved the model pays attention to physiologically relevant cardiac regions and phases. The framework maintained robust generalizability across imaging protocols and institutions with minimal performance degradation. By synergizing biomechanical insights with deep learning, our approach offers an interpretable, data-efficient solution for early NICM detection and subtype differentiation, holding strong translational potential for clinical practice. Full article

Over the past four decades, Wadi Fatimah in western Saudi Arabia has undergone significant environmental changes that have contributed to desertification. High-resolution spatial and temporal analyses are essential for monitoring the extent of desertification and understanding its driving factors. This study aimed to assess the spatial distribution of desertification in Wadi Fatimah using satellite and climate data. Landsat imagery from 1984 to 2022 was employed to derive land surface temperature (LST) and assess vegetation trends using the Normalized Difference Vegetation Index (NDVI). Climate variables, including precipitation and evapotranspiration (ET), were sourced from the gridded TerraClimate dataset (1980–2022). LST estimates were validated using MOD11A2 products (2001–2022), while TerraClimate precipitation data were evaluated against observations from four local rain gauge stations: Wadi Muharam, Al-Seal Al-Kabeer, Makkah, and Baharah Al-Jadeedah. A Desertification Index (DI) was developed based on four variables: NDVI, LST, precipitation, and ET. Five regression models—ridge, lasso, elastic net, polynomial regression (degree 2), and random forest regression—were applied to evaluate the predictive capacity of these variables in explaining desertification dynamics. Among these, Random Forest and Polynomial Regression demonstrated superior predictive performance. The classification accuracy of the desertification map showed high overall accuracy and a strong Kappa coefficient. Results revealed extensive land degradation in the central and lower sub-basins of Wadi Fatimah, driven by both climatic stressors and anthropogenic pressures. LST exhibited a clear upward trend between 1984 and 2022, especially in the lower sub-basin. Precipitation and ET analysis confirmed the region’s arid climate, characterized by limited rainfall and high ET, which exacerbate vegetation stress and soil moisture deficits. Validation of LST with MOD11A2 data showed reasonable agreement, with RMSE values ranging from 2 °C to 6 °C and strong correlation coefficients across most years. Precipitation validation revealed low correlation at Al-Seal Al-Kabeer, moderate at Baharah Al-Jadeedah, and high correlations at Wadi Muharam and Makkah stations. These results highlight the importance of developing robust validation methods for gridded climate datasets, especially in data-sparse regions. Promoting sustainable land management and implementing targeted interventions are vital to mitigating desertification and preserving the environmental integrity of Wadi Fatimah. Full article

Unconventional reservoirs are characterized by intricate fluid-phase behaviors, and physics-based shale oil well productivity prediction models often exhibit substantial deviations due to oversimplified theoretical frameworks and challenges in parameter acquisition. Under these circumstances, data-driven approaches leveraging actual production datasets have emerged as viable alternatives for productivity forecasting. Nevertheless, conventional data-driven architectures suffer from structural simplicity, limited capacity for processing low-dimensional feature spaces, and exclusive applicability to intra-sequence learning paradigms (e.g., production-to-production sequence mapping). This fundamentally conflicts with the underlying principles of mechanistic modeling, which emphasize pressure-to-production sequence transformations. To address these limitations, we propose a hybrid deep learning architecture integrating convolutional neural networks with bidirectional gated recurrent units (CNN-BiGRU). The model incorporates dedicated input pathways: fully connected layers for feature embedding and convolutional operations for high-dimensional feature extraction. By implementing a sequence-to-sequence (seq2seq) architecture with encoder–decoder mechanisms, our framework enables cross-domain sequence learning, effectively bridging pressure dynamics with production profiles. The CNN-BiGRU model was implemented on the TensorFlow framework, with rigorous validation of model robustness and systematic evaluation of feature importance. Hyperparameter optimization via grid searching yielded optimal configurations, while field applications demonstrated operational feasibility. Comparative analysis revealed a mean relative error (MRE) of 16.11% between predicted and observed production values, substantiating the model’s predictive competence. This methodology establishes a novel paradigm for machine learning-driven productivity prediction in unconventional reservoir engineering. Full article

Historical archives hold untapped potential for understanding long-term biodiversity change. This study introduces computational approaches to historical ecology, combining archival research, text analysis, and spatial mapping to reconstruct past biodiversity patterns. Using the 1845 Bavarian Animal Observation Dataset (AOD1845), a comprehensive survey of vertebrate species across 119 districts, we transform 5400 prose records into structured ecological data. Our analyses reveal how species distributions, habitat associations, and human–wildlife interactions were shaped by land use and environmental pressures in pre-industrial Bavaria. Beyond documenting ecological baselines, the study captures early perceptions of habitat loss and species decline. We emphasise the critical role of historical expertise in interpreting archival sources and avoiding anachronisms when integrating historical data with modern biodiversity frameworks. By bridging the humanities and environmental sciences, this work shows how digitised archives and computational methods can open new frontiers for conservation science, restoration ecology, and Anthropocene studies. The findings advocate for the systematic mobilisation of historical datasets to better understand biodiversity change over time. Full article

Ensuring the overall efficiency of hydraulic fracturing treatment depends on the ability to forecast bottomhole pressure. It has a direct impact on fracture geometry, production efficiency, and cost control. Since the complications present in contemporary operations have proven insufficient to overcome inherent uncertainty, the precision of bottomhole pressure predictions is of great importance. Achieving this objective is possible by employing machine learning algorithms that enable real-time forecasting of bottomhole pressure. The primary objective of this study is to produce sophisticated machine learning algorithms that can accurately predict bottomhole pressure while injecting guar cross-linked fluids into the fracture string. Using a large body of work, including 42 vertical wells, an extensive dataset was constructed and meticulously packed using processes such as feature selection and data manipulation. Eleven machine learning models were then developed using parameters typically available during hydraulic fracturing operations as input variables, including surface pressure, slurry flow rate, surface proppant concentration, tubing inside diameter, pressure gauge depth, gel load, proppant size, and specific gravity. These models were trained using actual bottomhole pressure data (measured) from deployed memory gauges. For this study, we carefully developed machine learning algorithms such as gradient boosting, AdaBoost, random forest, support vector machines, decision trees, k-nearest neighbor, linear regression, neural networks, and stochastic gradient descent. The MSE and R² values of the best-performing machine learning predictors, primarily gradient boosting, decision trees, and neural network (L-BFGS) models, demonstrate a very low MSE value and high R² correlation coefficients when mapping the predictions of bottomhole pressure to actual downhole gauge measurements. R² values are reported as 0.931, 0.903, and 0.901, and MSE values are reported at 0.003, 0.004, and 0.004, respectively. Such low MSE values together with high R² values demonstrate the exceptionally high accuracy of the developed models. By illustrating how machine learning models for predicting pressure can act as a viable alternative to expensive downhole pressure gauges and the inaccuracy of conventional models and correlations, this work provides novel insight. Additionally, machine learning models excel over traditional models because they can accommodate a diverse set of cross-linked fracture fluid systems, proppant specifications, and tubing configurations that have previously been intractable within a single conventional correlation or model. Full article

Eurasian drylands are vital for the global climate and ecological balance. Quantifying spatiotemporal variations in surface soil moisture (SSM) is essential for monitoring water, energy, and carbon cycles. The suitability of recent global-scale surface soil moisture datasets for Eurasian arid and semi-arid regions has not been comprehensively evaluated. This study investigates spatiotemporal trends of five SSM products—MERRA-2, ESACCI, GLEAM, GLDAS, and ERA5—from 1980 to 2023. The performance of these products was evaluated using in situ station data and the three-cornered hat (TCH) method, followed by partial correlation analysis to assess the influence of environmental factors, including mean annual temperature (MAT), mean annual precipitation (MAP), potential evapotranspiration (PET), vapor pressure deficit (VPD), and leaf area index (LAI), on SSM from 1981 to 2018. The results showed consistent SSM patterns: higher values in India, the North China Plain, and Russia, and lower values in the Arabian Peninsula, the Iranian Plateau, and Central Asia. Regionally, MAT, PET, VPD, and LAI increased significantly (0.04 °C yr⁻¹, 1.66 mm yr⁻¹, 0.004 kPa yr⁻¹, and 0.003 m² m⁻² yr⁻¹, respectively; p < 0.05), while MAP rose non-significantly (0.29 mm yr⁻¹). ERA5 exhibited the strongest correlation with in situ station data (R² = 0.42), followed by GLEAM (0.37), ESACCI (0.28), MERRA2 (0.19), and GLDAS (0.17). Additionally, ERA5 showed the highest correlation (correlation = 0.72), while GLEAM had the lowest bias (0.03 m³ m⁻³) and ESACCI exhibited the lowest ubRMSE (0.03 m³ m⁻³). The three-cornered hat method identified ERA5 and GLDAS as having the lowest uncertainties (<0.03 m³ m⁻³), with ESACCI exceeding 0.05 m³ m⁻³ in northern regions. Across land cover types, cropland had the lowest uncertainty among the five SSM products, while forest had the highest. Partial correlation and dominant factor analysis identified MAP as the primary driver of SSM. This study comprehensively evaluated SSM products, highlighting their strengths and limitations. It underscored MAP’s crucial role in SSM dynamics and provided insights for improving SSM datasets and water resource management in drylands, with broader implications for understanding the hydrological impacts of climate change. Full article

Pacific crown-of-thorns starfish (Acanthaster solaris) outbreaks pose a significant threat to coral reef ecosystems, with climate change potentially exacerbating their distribution and impact. However, there remains only a small number of predictive studies on how climate change drives changes in the distribution patterns of A. solaris, and relevant assessments of the impact of these changes on coral reef areas are lacking. To address this issue, this study investigated potential changes in the distribution of A. solaris under climate change and its impact on Acropora coral habitats. Using a novel two-step framework, we integrated both abiotic and biological (Acropora distribution) predictors into species distribution modeling to project future shifts in A. solaris habitats. We created the first reliable set of current and future global distribution maps for A. solaris using a comprehensive dataset and machine learning approach. The results showed significant distribution shifts under three climate change scenarios (SSP1-2.6, SSP2-4.5, and SSP5-8.5), with expanded ranges under all scenarios, and the greatest expansion occurring near 10° S. Asymmetry in the latitudinal shifts in habitat boundaries suggests that the Southern Hemisphere may face a more severe expansion of A. solaris. Regions previously unsuitable for A. solaris, such as parts of New Zealand, might experience new invasions. Additionally, our findings highlight the potential increase in predatory pressure on coral reefs under SSP2-4.5 and SSP5-8.5 scenarios, particularly in the Western Coral Triangle and Northeast Australian Shelf, where an overlap between A. solaris and Acropora habitats is significant. This study provides critical insights into the ecological dynamics of A. solaris in the context of climate change, and the results have important implications for coral reef management. These findings highlight the need for targeted conservation efforts and the development of mitigation strategies to protect coral reefs from the growing threat posed by A. solaris. Full article

Understanding the characteristics of wildfires in North China is critical for advancing regional fire danger prediction and management strategies. This study employed satellite-based burned area products of the Global Fire Emissions Database (GFED) and reanalysis of climate datasets to investigate the spatiotemporal characteristics of wildfires, as well as their relationships with fire danger indices and climatic drivers. The results revealed distinct seasonal variability, with the maximum burned area extent and intensity occurring during the March–April period. Notably, the fine fuel moisture code (FFMC) demonstrated a stronger correlation with burned areas compared to other fire danger or climate indices, both in temporal series and spatial patterns. Further analysis through the self-organizing map (SOM) clustering of FFMC composites then revealed six distinct modes, with the SOM1 mode closely matching the spatial distribution of burned areas in North China. A trend analysis indicated a 7.75% 10a⁻¹ (p < 0.05) increase in SOM1 occurrence frequency, associated with persistent high-pressure systems that suppress convective activity through (1) inhibited meridional water vapor transport and (2) reduced cloud condensation nuclei formation. These synoptic conditions created favorable conditions for the occurrence of wildfires. Finally, we developed a prediction model for burned areas, leveraging the strong correlation between the FFMC and burned areas. Both the SSP245 and SSP585 scenarios suggest an accelerated, increasing trend of burned areas in the future. These findings emphasize the importance of understanding the spatiotemporal characteristics and underlying causes of wildfires, providing critical insights for developing adaptive wildfire management frameworks in North China. Full article

In elevator systems, pressure plates secure guide rails and limit displacement, but defects compromise their performance under stress. Current detection algorithms face challenges in achieving high localization accuracy and computational efficiency when detecting small defects in guide rail pressure plates. To overcome these limitations, this paper proposes a lightweight defect detection network (LGR-Net) for guide rail pressure plates based on the YOLOv8n algorithm. To solve the problem of excessive model parameters in the original algorithm, we enhance the baseline model’s backbone network by incorporating the lightweight MobileNetV3 and optimize the neck network using the Ghost convolution module (GhostConv). To improve the localization accuracy for small defects, we add a high-resolution small object detection layer (P2 layer) and integrate the Convolutional Block Attention Module (CBAM) to construct a four-scale feature fusion network. This study employs various data augmentation methods to construct a custom dataset for guide rail pressure plate defect detection. The experimental results show that LGR-Net outperforms other YOLO-series models in terms of overall performance, achieving optimal results in terms of precision (p = 98.7%), recall (R = 98.9%), mAP (99.4%), and parameter count (2,412,118). LGR-Net achieves low computational complexity and high detection accuracy, providing an efficient and effective solution for defect detection in elevator guide rail pressure plates. Full article

Floristic and biodiversity knowledge play a crucial role in ecosystem conservation and sustainable land management, particularly in urban-rural contexts that can serve as biodiversity reservoirs, hosting species of high biogeographic value. Focusing on the new municipality of Misiliscemi, established in 2021 in Sicily and now facing the challenge of developing new management strategies, this study provides fundamental knowledge on the plant biodiversity of the area and explores how the integration of floristic and environmental data can guide territorial planning strategies aimed at preserving natural capital and ecosystem services. The research, based on field surveys conducted over many years, taxonomic identification of species, analysis of biological forms and chorological data, evaluation of ecological indicators, and GIS-based habitat mapping according to the EUNIS classification, has made it possible to obtain a comprehensive dataset. The results of this work led to the identification of 623 taxa, recording new findings for the Sicilian flora, including both native and alien species, which represent primary biodiversity data crucial for plant resource management. In addition, 42 habitat types were mapped, highlighting that approximately 80% of the territory is occupied by vegetated man-made habitats. Despite anthropogenic pressures and landscape modifications, Misiliscemi retains significant plant biodiversity, including habitats and species of conservation interest, that represent a vital resource for natural capital and ecosystem services. This knowledge base, in addition to constituting the scientific foundation upon which this young municipality can develop an urban planning strategy aimed at achieving sustainable local development, also represents a methodological approach that highlights how basic knowledge of urban biodiversity should be considered a crucial aspect of sustainable urban planning worldwide. Full article