Local Patterns Research Articles

Characterizing humpback whale behavior along the North-Norwegian coast

BackgroundStudying movement patterns of individual animals over time can give insight into how they interact with the environment and optimize foraging strategies. Humpback whales (Megaptera novaeangliae) undertake long seasonal migrations between feeding areas in polar regions and breeding grounds in tropical areas. During the last decade, several individuals have had up to a 3-month stop-over period around specific fjord-areas in Northern Norway to feed on Norwegian spring-spawning (NSS-) herring (Culpea harengus L.). Their behavioral patterns during this period are not well understood, including why some whales seemingly leave the fjords and then later return within the same season.MethodsTo investigate whale behavior during this seasonal stopover, we classified humpback whale tracks into five distinct movement modes; ranging, encamped, nomadic, roundtrip and semi-roundtrip. A behavioral change point analysis (BCPA) was used to select homogeneous segments based on persistence velocity. Then, net squared displacement (NSD) over time was modeled to differentiate movement modes. This study also manually identified longer roundtrips away from the fjords that lasted several days and examined movement modes within these.ResultsInside the fjord systems, encamped mode was most prevalent in December–January, suggesting the whales were mainly foraging on overwintering NSS-herring in this area. During the same winter seasons, half of the whales left the fjords and then returned. We hypothesize that these trips serve as ‘searching trips’ during which the whales seek better feeding opportunities outside the fjords. If better foraging conditions are not found, they return to the fjords to continue their feeding. The overall most common mode was ranging (54%), particularly seen during the start of their southwards migration and in areas outside the fjord systems, indicating that the whales mainly moved over larger distances in the offshore habitat.ConclusionsThis study serves as a baseline for future studies investigating both the searching trip theory and humpback whale behavior in general, and confirms that this method could be useful to analyze local scale movement patterns of satellite tagged whales.

Local patterns of spread of influenza A H3N2 virus in coastal Kenya over a 1-year period revealed through virus sequence data.

The patterns of spread of influenza A viruses in local populations in tropical and sub-tropical regions are unclear due to sparsity of representative spatiotemporal sequence data. We sequenced and analyzed 58 influenza A(H3N2) virus genomes sampled between December 2015 and December 2016 from nine health facilities within the Kilifi Health and Demographic Surveillance System (KHDSS), a predominantly rural region, covering approximately 891km2 along the Kenyan coastline. The genomes were compared with 1571 contemporaneous global sequences from 75 countries. We observed at least five independent introductions of A(H3N2) viruses into the region during the one-year period, with the importations originating from Africa, Europe, and North America. We also inferred 23 virus location transition events between the nine facilities included in the study. International virus imports into the study area were captured at the facilities of Chasimba, Matsangoni, Mtondia, and Mavueni, while all four exports from the region were captured from the Chasimba facility, all occurring to Africa destinations. A strong spatial clustering of virus strains at all locations was observed associated with local evolution. Our study shows that influenza A(H3N2) virus epidemics in local populations appear to be characterized by limited introductions followed by significant local spread and evolution. Knowledge of the viral lineages that circulate within specific populations in understudied tropical and subtropical regions is required to understand the full diversity and global ecology of influenza viruses and to inform vaccination strategies within these populations.

The Global and Local Patterns of Reading-Related Cognitive and Ecological Variables in Chinese First-Grade Children: A Cross-Sectional Network Analysis.

In the current study, we tested a network model of reading difficulty by using state-of-the-art psychological network analysis. Four hundred and fifty-three Chinese first-grade children (about 38% female, mean age = 7.00, SD = 0.41) were divided into good (n = 154), competent (n = 147), and struggling readers (n = 152) based on their scores of Chinese character reading. The Extended Bayesian Information Criterion graphical lasso (EBICglasso) method was applied to estimate cross-sectional networks for the three groups. Each network included four cognitive nodes (homophone awareness, morphological structure awareness, phonological awareness, and vocabulary) and two ecological nodes (family socioeconomic status and the number of books at home). Chronological age and nonverbal intelligence were also included in the estimated networks. The global (i.e., global structure and global connectivity) and local patterns (i.e., the most important edges and nodes) in each network were reported. The network comparison results showed that global connectivity was significantly lower among struggling readers than for good readers, implying that a holistic impairment of bidirectional connections among multiple variables relates to the difficulty in learning to read. The theoretical and empirical implications and the significance of applying the network approach to reading research are discussed.

Innovations through crop switching happen on the diverse margins of US agriculture

Crop switching, in which farmers grow a crop that is novel to a given field, can help agricultural systems adapt to changing environmental, cultural, and market forces. Yet while regional crop production trends receive significant attention, relatively little is known about the local-scale crop switching that underlies these macrotrends. We characterized local crop-switching patterns across the United States using the US Department of Agriculture (USDA) Cropland Data Layer, an annual time series of high resolution (30 m pixel size) remote-sensed cropland data from 2008 to 2022. We found that at multiple spatial scales, crop switching was most common in sparsely cultivated landscapes and in landscapes with high crop diversity, whereas it was low in homogeneous, highly agricultural areas such as the Midwestern corn belt, suggesting a number of potential social and economic mechanisms influencing farmers' crop choices. Crop-switching rates were high overall, occurring on more than 6% of all US cropland in the average year. Applying a framework that classified crop switches based on their temporal novelty (crop introduction versus discontinuation), spatial novelty (locally divergent versus convergent switching), and categorical novelty (transformative versus incremental switching), we found distinct spatial patterns for these three novelty dimensions, indicating a dynamic and multifaceted set of cropping changes across US farms. Collectively, these results suggest that innovation through crop switching is playing out very differently in various parts of the country, with potentially significant implications for the resilience of agricultural systems to changes in climate and other systemic trends.

Exploring transmission dynamics of the Sarcoptes scabiei mite in humans by combining molecular typing and epidemiological variables, the Netherlands 2016–2023

BackgroundScabies, an infestation of the mite Sarcoptes scabiei, has seen an increase in clinical diagnoses in the Netherlands since 2011. This study aimed to analyse PCR-positive S. scabiei skin samples through partial genome sequencing and to link findings to patient epidemiological characteristics.MethodsSkin samples were collected from individuals in the Netherlands between January 2016 and January 2023. On the PCR-positive S. scabiei skin samples, partial mitochondrial cytochrome c oxidase subunit 1 gene (cox1) sequencing was performed to assess genetic variability. Epidemiological information was collected through interviews. We examined associations between cox1 subtypes, epidemiological factors and treatment outcomes.ResultsSequencing results were obtained from 128 patients, with epidemiological information available for 55 (43%) of these patients. Fifteen distinct cox1 subtypes were identified. Subtype 01 was most prevalent (45%) and present across all age groups and social settings. The remaining subtypes were less common and not consistently found in all contexts. Five clusters were identified, each with identical cox1 subtypes. Comparative analysis with GenBank sequences revealed genetic similarities with strains from Australia, the USA and China, suggesting the global distribution and transmission of specific subtypes. A substantial proportion (73%) of patients with scabies required multiple treatments to eradicate the infestation, with no subtype-related differences.ConclusionsThis is the first study linking S. scabiei sequencing results to patient epidemiological data. Several subtypes clustered in specific geographic regions and social contexts, underscoring localised transmission patterns. Further research with larger sample sizes is needed to enhance our understanding of the transmission of this mite. This study provides valuable insights that will strengthen scabies control efforts.Graphical

Regularity and correlation analysis of regional lymph node metastasis in nonoperative patients with non-small cell lung cancer based on positron emission tomography/computed tomography images

BackgroundDefinitive concurrent chemoradiotherapy (CCRT) is the standard treatment for locally advanced, inoperable non-small cell lung cancer (NSCLC). Previous studies have mainly focused on examining local failure and recurrence patterns after surgery and the principles of lymph node metastasis (LNM) in surgical candidates with NSCLC. However, these studies were just only able to guide postoperative radiotherapy (PORT) and the patterns of LNM in patients with resected NSCLC was inadequate to represent that in locally advanced inoperable NSCLC patients for guiding target volume delineation of CCRT. In this study, we aimed to analyze the metastasis regularities and establish the correlations between different lymph node levels in NSCLC patients without any intervention using positron emission tomography/computed tomography (PET/CT) images.MethodsOverall, 358 patients with N1–N3 NSCLC admitted in our hospital between 2018 and 2022 were retrospectively analyzed. The diagnosis of metastatic lymph nodes was reviewed and determined using the European Organization for Research and Treatment of Cancer standard and the standardized value of the PET/CT examination. Univariate and multivariate analysis were performed to investigate the correlations between the different levels were evaluated by using of the chi-square test and logistic regression model.ResultsThe lymph nodes with the highest metastasis rates in patients with left lung cancer were in order as follows: 10L, 4L, 5, 4R, and 7; while in those with right lung cancer they were 10R, 4R, 7, 2R, and 1R. Notably, we found left lung patients were more likely to have contralateral hilar, mediastinal and supraclavicular lymph nodes involved, and the right lung group exhibited a higher propensity for ipsilateral mediastinum and supraclavicular lymph node invasion. Furthermore, correlation analysis revealed there were significant correlative patterns in the LNM across different levels.ConclusionsThis study elucidated the patterns of primary LNM in patients with NSCLC who had not undergone surgery (without any treatment interventions) and the correlations between lymph node levels. These findings were expected to provide useful reference for target volume delineation in definitive concurrent chemoradiotherapy in locally advanced NSCLC patients.

Analysis of spatial structure and singularity of microbial biogeochemical anomalies in the Taiwan Strait Basin

The abundance of specific hydrocarbon-oxidizing bacteria (MV value) in surface soil is closely linked to the concentration of thermogenic light hydrocarbons found in the underlying layers. The identification of these bacteria using specific detection techniques offers unique insights into microbiological anomalies associated with oil and gas, underscoring the significance of delineating microbiological geochemical anomalies in the exploration process. Accurate determination of the anomaly range is crucial for forecasting potential oil and gas reservoirs. In this study, a variety of fractal and geological statistical methods were employed to analyze microbiological geochemical data from the Taiwan Strait Basin. The multifractal approach allows for the simultaneous assessment of spatial autocorrelation and singularity in geochemical fields, with the fractal distribution reflecting the localized enrichment and depletion patterns of microbiological geochemical elements in rocks and other substrates. Geological statistical methods, such as the variation function, enable the assessment of spatial autocorrelation in geochemical and geophysical fields. The results indicate that microbiological anomalies in the Jiulongjiang Depression are mainly concentrated in the southwestern part of the study area, with some anomalies also observed in the central region. Conversely, microbiological anomalies in the Jinjiang Depression are predominantly located in the southern portion of the study area. Microbiological anomalies in the study area (Jinjiang Depression and Jiulongjiang Depression) exhibit a northeast-southwest strip distribution, aligning with the trend of the depressions and faults, suggesting that faults could influence the distribution of oil and gas resources in the region. Based on microbiological statistical analyses, the Jiulongjiang Depression shows better potential for oil and gas prospects compared to the Jinjiang Depression, guiding future research efforts in the Jiulongjiang Depression.

Optimised GRS coded cooperation: sliding relay selection design and improved joint decoding

ABSTRACT In this paper, the novel optimised generalised Reed-Solomon coded cooperative (OGRSCC) scheme is developed with a flexible selection of the code length and information dimension. In the OGRSCC scheme, two GRS codes with different dimensions are deployed at the source and relay, respectively, where the information symbols of the relay are dependent on those of the source. At the relay, different symbol selections may contribute to distinct joint codes at the destination. As a result, to achieve a maximum free distance (FD) and the optimised weight distribution of the final code, the global search method is first proposed to obtain the optimal selection pattern. However, as the code grows longer, its complexity rises dramatically. To make the selection more practical, the low-complexity sliding search method is then investigated to obtain the local optimal symbol selection pattern. Monte-Carlo simulated results indicate the OGRSCC scheme with two selection methods demonstrates almost identical performance. Furthermore, based on this special selection method and the unique algebraic structure of GRS code, the improved joint decoding algorithm is first proposed to boost the overall system performance. In addition, the theoretical upper bound is analysed, verifying the effectiveness of the proposed OGRSCC scheme. Numerical simulations reveal that the proposed OGRSCC scheme outperforms its corresponding non-cooperative scheme by a gain of over 3.2 dB and its existing comparable schemes by a margin of approximately more than 4 dB at high signal-to-noise ratio (SNR), respectively.

Kitaev honeycomb antiferromagnet in a field: quantum phase diagram for general spin

We use tensor-network methods and high-order linked-cluster expansions to explore the quantum phase diagram of the antiferromagnetic Kitaev honeycomb model in a magnetic field for general spin S values. Tensor network calculations for the pure Kitaev model confirm the absence of fluxes and spin-spin correlations beyond nearest neighbors, while revealing discrete orientational symmetry breaking for S ∈ 1, 3/2, 2, consistent with the semiclassical limit. An intermediate region between Kitaev phases and the high-field polarized phase is identified for all considered spin values, showing a sequence of potential phases characterized by distinct local magnetization patterns while the total magnetization increases smoothly as a function of the field. Linked-cluster expansions for the high-field zero-momentum gap and spectral weight indicate a quantum critical breakdown of the polarized phase, suggesting exotic physics at intermediate Kitaev couplings.

Detection of Dominant Rainfall Patterns in Indonesian Regions Using Empirical Orthogonal Function (EOF) and Its Relation with ENSO and IOD Events

Several studies indicate that Indonesia’s position on the equator, where winds from the northern and southern hemispheres converge, leads to year-round rainfall in Indonesia. Previous studies have also shown the presence of monsoon winds over Indonesia, which causes rain in Indonesia to follow the Asian and Australian monsoon wind patterns. Recently, variations in rainfall patterns and intensity in Indonesia have been highly volatile and difficult to predict. This study uses gridded data analysis of ocean-atmospheric rainfall for 69 years (1948–2016) to detect monsoonal, equatorial, and local signals and relate them to ENSO and IOD events. Rainfall analysis reveals that the first mode accounts for 38.57%, the second 13.92%, and the third 8.93% of the total variance. These results show that these variants are very variable, both in space and time. Monsoonal patterns were detected in the southern region of Indonesia, equatorial patterns in northern Sumatra and western Kalimantan, while local patterns were detected in Maluku and the Maluku Islands. Furthermore, the monsoonal and local rainfall patterns are strongly correlated (> 0.85) with local SST, but the equatorial rainfall pattern is weakly correlated (< 0.5) with local SST. Note that the equatorial pattern is associated with the Inter Tropical Convergence Zone (ITCZ) phase. The varied vector correlation map illustrates the asymmetric fluctuations in sea surface temperature and the simultaneous existence of winds coming from the west from the Indian Ocean and winds coming from the east from the Pacific Ocean during the period of heavy precipitation.

X-Band Radar and Surface-Based Observations of Cold-Season Precipitation in Western Colorado’s Complex Terrain

Abstract Hydrologic processes associated with intermountain cold-season precipitation in the Upper Colorado River basin have important impacts on avalanche forecasting and water resource management. However, traditional weather radar networks struggle with observations in this complex terrain. Data collected during the Study of Precipitation, the Lower Atmosphere, and the Surface for Hydrometeorology (SPLASH) and its sister campaign, Surface Atmosphere Integrated Field Laboratory (SAIL) in the East River watershed of western Colorado, are used to examine a multistorm period from 23 December 2021 to 1 January 2022 that contributed 35% of the total winter precipitation in this watershed. Dual-polarization X-band radar and disdrometer measurements show ∼30-mm differences in precipitation amount at two sites in proximity over four distinct storm events within the period. Wind patterns, synoptic forcings, microphysical characteristics of precipitation, and surface meteorology are analyzed to explain the observed spatial variability of cold-season precipitation in complex mountainous terrain. Analysis shows that differences over time within this event are mainly accounted for by synoptic forcings, such as frontal passages; differences between sites are accounted for by the impact of variations in local wind patterns on precipitation microphysics. Patterns of surface precipitation intensity are compared and found to be correlated with X-band radar signatures; a relationship between a strong dendritic growth stage and intense low-density surface precipitation is reinforced by this study. This relationship demonstrates the importance of particle growth mechanisms on surface snowfall patterns in high-altitude complex terrain, underscoring the importance of realistic microphysical parameterizations. Significance Statement The amount and density of snowpack from western Colorado winter storms have significant impacts on water resources in the Upper Colorado River basin. Snowpack characteristics are affected by small-scale differences in how snow forms in the atmosphere. These differences are hard to study in the complex terrain of the Rockies, but data from the SPLASH and SAIL field campaigns allows us to investigate how snow crystal formation and mountain-driven wind patterns affect snow near the surface. Our study finds that snow crystal growth varies over small space and time scales and is likely controlled by the terrain beneath a given location and resultant local wind patterns. These results imply that predicting snowpack in the Rockies requires properly representing local wind patterns and crystal growth processes in models.

Evolution process of chemical weathering and sediment sources in the Makran Continental margin since the Younger Dryas

The chemical weathering processes and sedimentary source evolution since the Younger Dryas (YD) in the low-latitude arid continental margin have been investigated. Two sediment cores, MK07G and MK09G, were retrieved from the Makran continental margin in the northern Arabian Sea and subjected to analyses of major and trace elements, along with AMS14C dating. The results show that since the YD, the weathered parent rocks of Makran sediments have remained relatively stable, predominantly consisting of felsic rocks, with some contributions from mafic rocks. The Makran sediments exhibit initial to moderate weathering, with no discernible effects from grain size sorting or disturbances from sediment recycling, indicating primary deposition. Significant contributions of terrigenous eolian dust from surrounding continents (e.g., the Indian subcontinent, Arabian Peninsula, and northeastern Africa) were identified, along with riverine inputs from the Dasht River and fine-grained components from the Late Pleistocene Indus delta sediment, as well as proximal basin sedimentation. The evolution of sediment sources in the study area is significantly influenced by the Indian Monsoon and westerly wind systems, with intensified monsoon phases and westerly conditions correlating with increased fluvial input. Furthermore, chemical weathering processes since the YD are closely linked to local precipitation patterns, where intensified rainfall enhances weathering intensity. Records from the Makran continental margin indicate a teleconnection between chemical weathering and sedimentary processes in the Arabian Sea and Bond events in the North Atlantic, highlighting the extensive influence of Northern Hemisphere climate fluctuations.

A bridge to clean cooking? The cost-effectiveness of energy-efficient biomass stoves in rural Senegal

Rural areas of sub-Saharan Africa have experienced limited progress towards the sustainable development goal of universal access to clean cooking. Energy-efficient biomass cookstoves (EEBCs) are considered a potential bridge technology, but EEBC models vary widely, and there is a lack of understanding about their real-world use implications. We conduct a randomized controlled trial in rural Senegal to compare a low-cost, locally produced stove designed to achieve fuel savings and an expensive, imported stove shown to be more efficient and emissions-reducing in the laboratory. We find that the two EEBCs perform similarly: both reduce fuel consumption but have no significant impact on cooking time and fuel collection, emissions, or objective health measures. We conclude that the technically advanced option is not cost effective for most of our sample, while the low-cost EEBC can be seen as a stop-gap solution that primarily reduces fuel use. The findings underpin the importance of customizing EEBC dissemination to local context and baseline cooking patterns.

Firefly forest: A swarm iteration-free swarm intelligence clustering algorithm

The Firefly Forest algorithm is a novel bio-inspired clustering method designed to address key challenges in traditional clustering techniques, such as the need to set a fixed number of neighbors, predefine cluster numbers, and rely on computationally intensive swarm iterative processes. The algorithm begins by using an adaptive neighbor estimation, refined to filter outliers, to determine the brightness of each firefly. This brightness guides the formation of firefly trees, which are then merged into cohesive firefly forests, completing the clustering process. This approach allows the algorithm to dynamically capture both local and global patterns, eliminate the need for predefined cluster numbers, and operate with low computational complexity. Experiments involving 14 established clustering algorithms across 19 diverse datasets, using 8 evaluative metrics, demonstrate the Firefly Forest algorithm’s superior accuracy and robustness. These results highlight its potential as a powerful tool for real-world clustering applications. Our code will be open-sourced after the peer review.

The ecological and evolutionary dynamics of inselbergs.

Islands are fundamental model systems in ecology, biogeography, and evolutionary biology. However, terrestrial islands, unlike their aquatic counterparts, have received comparatively less attention. Among these land islands, inselbergs (i.e. isolated rock outcrops with diverse lithologies and a modest topographical prominence) stand out as iconic examples distributed worldwide across global biomes. Due to their durable lithology, inselbergs change slowly, persisting for tens of millions of years. In this review, we propose a biological definition for inselbergs that captures three fundamental characteristics of inselbergs from the perspective of biota. These are old age, isolation and the presence of unique microhabitats that are rare or absent in the surrounding matrix, fostering distinct communities often with unique and endemic biota. We synthesise the state of the art and formulate a set of testable hypotheses to deepen our understanding of the origins and maintenance of diversity on inselbergs, which are increasingly exposed to anthropogenic threats. By offering different habitats compared to the surrounding habitat matrix (e.g. moist microhabitats in dryland landscapes and xeric environments in humid tropical landscapes), inselbergs may allow specific lineages to thrive beyond their typical geographical limits. Particularly in drylands and degraded landscapes, inselbergs may not just provide different habitats but also act as ecological refuges or evolutionary refugia by providing a wider range of potential microhabitats than the surrounding matrix, enhancing resilience and promoting regional biodiversity. The central role of the matrix ensures that the ecological and evolutionary dynamics of inselbergs differ from those of true islands such as oceanic islands. Given that inselberg biota coexist within a terrestrial matrix, interactions between inselberg and matrix populations impact each other significantly. Over evolutionary timescales, matrix species may contract to inselberg refugia, preserving lineages while cycles of isolation and reconnection may drive speciation via a species pump. Although inselberg biodiversity has been studied predominantly from an island biogeography perspective, we argue that depending on the spatial scale, habitat specificity and mobility of the organisms considered, a range of different theories and paradigms can help explain the biogeography and local distribution patterns of different taxonomic and functional groups of inselberg species.

FDNet: Frequency filter enhanced dual LSTM network for wind power forecasting

The inherent volatility and intermittency of wind power present significant forecasting challenges, undermining the efficient integration of wind energy into the power grid. Existing methodologies, notably long short-term memory (LSTM) networks, encounter significant limitations due to their inefficiencies in processing long sequences, difficulties in capturing multi-scale temporal dynamics, and heightened sensitivity to noisy data, which can severely hamper model performance. To address these challenges, This paper proposes the frequency filter enhanced dual LSTM network (FDNet), a novel approach that directly addresses the constraints of the LSTM and improves the accuracy and stability of wind power forecasting. Specifically, FDNet employs the patching operation to divide the original time series into several sub-sequences, potentially boosting the computational efficiency. Furthermore, a specific frequency filter is designed and incorporated into FDNet, effectively reducing the influence of noise. Finally, a dual LSTM structure is employed, which enables FDNet to adeptly discover both short-term local temporal patterns and long-term global temporal patterns inherent in wind power data. Extensive experiments across three datasets demonstrate that FDNet significantly outperforms existing methods, achieving up to 11.0% reduction in mean absolute error and 8.1% in root mean squared error on the HL dataset, underscoring its effectiveness in wind power forecasting.

Coupled Modeling of Computational Fluid Dynamics and Granular Mechanics of Sand Production in Multiple Fluid Flow

Summary Sand production is a significant issue in oil and gas fields with poorly consolidated formations, often involving the multiphase flow of reservoir fluids and solid particles. The multiscale mechanisms of sand production, particularly fluid flow and particle movement, remain poorly understood. This study investigates these mechanisms using a coupled computational fluid dynamics and discrete element method (CFD-DEM) modeling approach. Single and multiple fluid flows of water and heavy oil were simulated with increasing fluid injection velocities, leading to different sand production patterns. The simulation results were compared with experimental results from a large cylindrical specimen of weak artificial sandstone under similar loading conditions. The multiphase conditions created various localized flow and deformation patterns that influenced both fluid and solid production, resulting in shorter transient sand production periods. Microstructures and phenomena such as fingering and water coning were observed, associated with a critical flow rate below which oil displacement was uniform and no water breakthrough occurred. Higher fluid injection velocities and fluid viscosities resulted in greater drag forces, leading to progressive damage zones and explaining the occurrence of single or multiple staged sand production events. The evolution of the microscopic granular structure was visualized under the effect of transient sand production.

On the use of kolmogorov–arnold networks for adapting wind numerical weather forecasts with explainability and interpretability: application to madeira international airport

Abstract This study examines the application of machine learning to enhance wind nowcasting by using a Kolmogorov-Arnold Network model to improve predictions from the Global Forecast System at Madeira International Airport, a site affected by complex terrain. The research addresses the limitations of traditional numerical weather prediction models, which often fail to accurately forecast localized wind patterns. Using the Kolmogorov-Arnold Network model led to a substantial reduction in wind speed and direction forecast errors, with a performance that reached a 48.5% improvement to the Global Forecast System 3 h nowcast, considering the mean squared error. A key outcome of this study comes from the model’s ability to generate mathematical formulas that provide insights into the physical and mathematical dynamics influencing local wind patterns and improve the transparency, explainability, and interpretability of the employed machine learning models for atmosphere modeling.

RFNet: Multivariate long sequence time-series forecasting based on recurrent representation and feature enhancement

Multivariate time series exhibit complex patterns and structures involving interactions among multiple variables and long-term temporal dependencies, making multivariate long sequence time series forecasting (MLSTF) exceptionally challenging. Despite significant progress in Transformer-based methods in the MLSTF domain, many models still rely on stacked encoder–decoder architectures to capture complex time series patterns. This leads to increased computational complexity and overlooks spatial pattern information in multivariate time series, thereby limiting the model’s performance. To address these challenges, we propose RFNet, a lightweight model based on recurrent representation and feature enhancement. We partition the time series into fixed-size subsequences to retain local contextual temporal pattern information and cross-variable spatial pattern information. The recurrent representation module employs gate attention mechanisms and memory units to capture local information of the subsequences and obtain long-term correlation information of the input sequence by integrating information from different memory units. Meanwhile, we utilize a shared multi-layer perceptron (MLP) to capture global pattern information of the input sequence. The feature enhancement module explicitly extracts complex spatial patterns in the time series by transforming the input sequence. We validate the performance of RFNet on ten real-world datasets. The results demonstrate an improvement of approximately 55.3% over state-of-the-art MLSTF models, highlighting its significant advantage in addressing multivariate long sequence time series forecasting problems.

Computer-aided diagnosis of liver cancer with improved SegNet and deep stacking ensemble model

Liver cancer is a leading cause of cancer-related deaths, often diagnosed at advanced stages due to reliance on traditional imaging methods. Existing computer-aided diagnosis systems struggle with noise, anatomical complexity, and ineffective feature integration, leading to inaccuracies in lesion segmentation and classification. By effectively addressing these challenges, the model aims to enhance early detection and assist clinicians in making informed decisions. Ultimately, this research seeks to contribute to more efficient and accurate liver cancer diagnosis. This paper presents a novel model for liver cancer classification, called SegNet-based Liver Cancer Classification via SqueezeNet (SgN-LCC-SqN). The model effectively executes liver cancer segmentation and classification through four key steps: preprocessing, segmentation, feature extraction, and classification. During preprocessing, Quadratic Mean Estimated Wiener Filtering (QMEWF) is utilized to minimize image noise. Segmentation divides the image into segments using Enhanced Feature Pyramid SegNet (EFP-SgN), which is essential for precise diagnosis. Feature extraction encompasses color features, Local Directional Pattern Variance, and Correlation Filtering-Local Gradient Increasing Pattern (CF-LGIP) features. The extracted features are then processed through an ensemble model, Deep Convolutional, Recurrent, Long Short Term Memory with SqueezeNet (DCR-LSTM-SqN), which includes Deep Convolutional Neural Network (DCNN), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), and Modified Loss Function in SqueezeNet (MLF-SqN) classifiers, sequentially analyzing the feature sets through DCNN, RNN, and LSTM before classification by MLF-SqN. The performance of the suggested DCR-LSTM-SqN model is evaluated over conventional methods for positive, negative and other metrics. The DCR-LSTM-SqN model consistently demonstrates superior accuracy, ranging from 0.947 to 0.984, across all training data percentages. Thus, the proposed model effectively segments liver lesions and classifies cancerous areas, demonstrating its potential as a valuable resource for clinicians to enhance the efficiency and accuracy of liver cancer diagnosis.