Deep Samples Research Articles

Применение технологии объединения методов разведочных работ для достижения экологического эффекта на месторождениях Каспия

The method of performing exploration works allowing to obtain an environmental effect in the Caspian Sea fields is described in the article. The research area includes the Rakushechno-Latitudinal and Khvalynsko-Sarmatian zones. Deposits of these zones differ in a complex geological structure. When creating a model of deposits a set of data is used: high-precision 3D seismic surveys on the study area and the results of drilling wells. Such detail of seismic exploration works allows to obtain high-quality results that are due to the high density of reflection points over the area and to the maximum use of various dynamic parameters of reflected waves. The results of such surveys are used in the calculation of reserves, justification of work on additional exploration and subsequent operation. Based on the results of seismic surveys a quantitative forecast of reservoir properties was carried out due to the inversion. For implementation seismic data was processed while preserving the true amplitudes of the seismic signal. Drilling wells were also the main source of information when performing work on calculating reserves. The complex structure and heterogeneity conditions of reservoir properties in area and section, the presence of discontinuous violations led to the use of the described technology, which involves the assessment of reserves and their transfer to the industrial category without testing in the column and causing inflow to the surface. This is done by means of a comprehensive accelerated diagnostics of productive formations for the assessment of reserves of the industrial category. Such a diagnosis provides for the study of productive formations based on the analysis and generalization of accumulated material on core, GIS, GDIS, GDK-OPK and PVT studies, without conducting tests in wells. With the help of these studies the filtration-capacitance properties of rocks, the nature of fluid saturation, contact positions, static levels, reservoir and bottom-hole pressures and reservoir temperatures, as well as deep sampling are determined. The use of modern instruments such as MDT (reservoir Tester) makes it possible to obtain reliable deep samples of the reservoir fluid. The results of data interpretation are successfully approved when making reports on UVS reserves calculation.

Pore connectivity and microfracture characteristics of Longmaxi shale in the Fuling gas field: Insights from mercury intrusion capillary pressure analysis

Assessing the connectivity of pores and the characteristics of microfractures is essential for comprehending gas migration mechanisms and effectively implementing development strategies in deep shale gas plays. The Fuling gas field, a prominent shale gas reservoir in China, holds promise for deep shale development. This requires a better understanding of variations in pore connectivity among shale samples from different depths of the Fuling gas field. To achieve this goal, this study employs innovative methodologies based on mercury intrusion capillary pressure (MICP), a recognized technique for evaluating material pore structures. In this research, we conducted MICP tests on shale samples of varying sizes to examine variations in pore structure and quantify inaccessible pores. Additionally, repeated MICP tests were performed on the same sample to investigate the spatial distribution of residual mercury and distinguish different pore types associated with flow conduits and storage spaces. Furthermore, a directional MICP experiment was carried out to quantitatively characterize the degree of pore development and anisotropy of microfractures by considering the direction of mercury intruding the sample (parallel or perpendicular to lamination). Our findings revealed that deep shale samples exhibited a higher proportion of inaccessible pores compared to mid-deep ones. Regarding accessible pores, a greater proportion of connected pores was observed in the deep shale samples, while the mid-deep shale samples exhibited a lower proportion. The results from the directional MICP experiment indicated that samples with anisotropy in mercury intrusion volume possessed finer and denser microfractures. However, no evidence was found to correlate the ratio of dead-end and connected pores with sample anisotropy. Overall, this study presents a novel approach for experimentally evaluating pore connectivity and microfracture characteristics in shale reservoirs using MICP. The outcomes contribute to enhancing our understanding of shale gas migration mechanisms, particularly in deep ultra-low-permeability reservoirs, and hold significant implications for future developments in the field.

Exploring paleoproterozoic metamorphism in the western Río de la Plata craton from drill-core evidence

Rhyacian basement is not exposed in the western part of the Río de la Plata craton, but deep borehole samples (depth of 1046–3340 m) provide information on its composition and metamorphic evolution. An amphibole schist sample from the Camilo Aldao borehole underwent two metamorphic stages: M1 and M2. The P-T conditions of M1 are not constrained whereas the M2 main metamorphic stage is characterized by conditions of 6.1–7.4 kbar and 575–605 °C. In contrast, an olivine gabbronorite sample from the Santiago Temple borehole records crystallization conditions of 1.7 ± 0.6 kbar and 1180 ± 10 °C based on clinopyroxene composition, and late magmatic to subsolidus low-pressure granulite facies metamorphism at 2.3 ± 1.2 kbar and 749 ± 85 °C based on multi-equilibrium thermobarometry. The Rhyacian evolution of the Río de la Plata craton is characterized by at least three thermo-tectonic orogenic stages: T1 (ca. 2200–2100 Ma), T2 (ca. 2100–2040 Ma), and T3 (ca. 2040–2010 Ma) that correspond respectively to: 1) early oceanic subduction and juvenile continental arc magmatism; 2) subsequent continental collision (clockwise intermediate P/T paths of metamorphism), and 3) final low-pressure metamorphism coeval with post-orogenic magmatism. The Camilo Aldao amphibole schist is representative of the T2 orogenic stage while the Santiago Temple gabbronorite is representative of T3 magmatism and metamorphism.

Bacterial Sulfate Reduction Facilitates Iodine Mobilization in the Deep Confined Aquifer of the North China Plain.

Bacterial sulfate reduction plays a crucial role in the mobilization of toxic substances in aquifers. However, the role of bacterial sulfate reduction on iodine mobilization in geogenic high-iodine groundwater systems has been unexplored. In this study, the enrichment of groundwater δ34SSO4 (15.56 to 69.31‰) and its significantly positive correlation with iodide and total iodine concentrations in deep groundwater samples of the North China Plain suggested that bacterial sulfate reduction participates in the mobilization of groundwater iodine. Similar significantly positive correlations were further observed between the concentrations of iodide and total iodine and the relative abundance of the dsrB gene by qPCR, as well as the composition and abundance of sulfate-reducing bacteria (SRB) predicted from 16S rRNA gene high-throughput sequencing data. Subsequent batch culture experiments by the SRB Desulfovibrio sp. B304 demonstrated that SRB could facilitate iodine mobilization through the enzyme-driven biotic and sulfide-driven abiotic reduction of iodate to iodide. In addition, the dehalogenation of organoiodine compounds by SRB and the reductive dissolution of iodine-bearing iron minerals by biogenic sulfide could liberate bound or adsorbed iodine into groundwater. The role of bacterial sulfate reduction in iodine mobilization revealed in this study provides new insights into our understanding of iodide enrichment in iodine-rich aquifers worldwide.

Deep attention sampling hashing for efficient image retrieval

Hashing has received broad attention in large-scale image retrieval due to its appealing efficiency in computation and storage. Particularly, with the drawn of deep learning, much efforts have been directed towards using deep neural networks to learn feature representations and hash codes simultaneously, and the developed deep hashing methods have shown superior performance over conventional hashing methods. In this paper, we propose Deep Attention Sampling Hashing (DASH), a novel deep hashing method that yields high-quality hash codes to enable efficient image retrieval. Specifically, we employ two sub-networks in DASH, i.e., a master branch and a part branch, to capture global structure features and discriminative feature representations, respectively. Furthermore, we develop an Attention Sampler Module (ASM), which consists of an Object Region Extraction (ORE) block and an Informative Patch Generation (IPG) block, to yield richer informative image patches. The ORE block provides a well-designed multi-scale attentional fusion mechanism to highlight and extract the significant regions of images, and the IPG block employs a direction-specific shift mechanism to generate desired image patches with discriminative details. Both blocks could be seamlessly integrated into various convolutional neural network (CNN) architectures. Subsequently, we conduct knowledge distillation optimization to transfer the details learned by the part branch into the master branch to guide hash code learning. In addition, we design a Weibull quantization loss to minimize the information loss caused by binary quantization. The experimental results on three benchmark datasets demonstrate the effectiveness of the proposed DASH with respect to different evaluation metrics.

Detection of Cutibacterium (Propionibacterium) acnes in orthopaedic surgery: serious problem or contamination?

PurposeBone and joint infections are an important and increasing problem. Whether intraoperatively detected bacteria should be considered relevant or not is often difficult to assess.This retrospective cohort study analyzes the relevance of C. acnes cultured from deep intraoperative specimens.MethodsAll deep tissue samples collected intraoperatively between 2015 and 2020 from a quartiary care provider were evaluated for detection of C. acnes and its therapeutical consequences. Infection rates were determined according to a standardized definition and protocol and analyzed in dependence of patient’s demographic data (age and gender), operative parameters (type of surgery, body region/location of surgery, and impression of the surgeon), and initiated therapy.ResultsIn 270 cases of more than 8500 samples, C. acnes was detected. In 30%, the detection was considered an infection. The number of samples taken and tested positive for C. acnes correlated significantly with its classification as a cause of infection. If more than one sample of the patient was positive, the detection was significantly more likely to be treated as infection (p < 0.001).In 76% of cases, a consultation to the infectious diseases (ID) department took place regarding the classification of the pathogen detection and the therapy to be carried out.Almost all of the tested isolates demonstrated the wild-type susceptibility for penicillin and clindamycin.ConclusionIntraoperative detection of skin-colonizing bacteria such as C. acnes is not always synonymous with infection. In particular, if other examination results contradict an infection (pathological sample without evidence of an infectious event, detection of malignant cells, etc.), the situation must be considered in a very differentiated manner. Interdisciplinary boards, for example, are suitable for this purpose. Care should be taken to obtain a sufficiently large number of tissue samples for microbiological examination to be able to better classify the result.

Soil geochemistry of hydrogen and other gases along the San Andreas fault

Natural hydrogen has generated great interest as a potential clean and renewable energy source. To understand the occurrence of natural hydrogen, 103 1-m deep soil gas samples were acquired near the San Andreas Fault at Jasper Ridge and Portola Valley, California, USA. The gas samples were analyzed for hydrogen, helium, carbon dioxide, light hydrocarbons, and fixed gas concentrations. Statistical data analysis was carried out to group samples, reveal their spatial distribution, and understand possible sources of the gases.High concentrations of hydrogen up to 20.3 ppmv and 17.3 ppmv occur in Jasper Ridge and Portola Valley, respectively, ∼30–35 times greater than the atmospheric concentration. Most samples with high hydrogen concentrations fall on or near faults, suggesting an origin by serpentinization or geomechanical activation of catalytic sites in minerals, although a deep-seated primordial origin cannot be excluded. Elevated concentrations of carbon dioxide resulted from aerobic microbial degradation of organic matter and elevated concentrations of light hydrocarbons likely resulted from thermal cracking of organic matter.

Building Brand Love for Travelers Through Accessibility, Destination Image, and Emotional Energy in Individuals at Laut Tawar Lake

Deep goals study This is for know directly influences accessibility, destination image, and emotional energy in individuals towards brand love. Type of research that is study quantitative. The population is previous tourists who have visited Lake Laut Tawar for at least two visits. Deep sample study This was taken with an accidental sampling technique of 100 tourists. Data was collected with a compiled questionnaire based on each indicator variable research. Analysis results study show that accessibility, destination image, and emotional energy in individuals direct influential significance towards brand love. The most dominant variable influencing tourist brand love is accessibility compared with destination image and emotional energy in individuals.

Novel RNA viruses from the Atlantic Ocean: Ecogenomics, biogeography, and total virioplankton mass contribution from surface to the deep ocean.

Marine viruses play a major role in the energy and nutrient cycle and affect the evolution of their hosts. Despite their importance, there is still little knowledge about RNA viruses. Here, we have explored the Atlantic Ocean, from surface to deep (4.296 m), and used viromics and quantitative methods to unveil the genomics, biogeography, and the mass contribution of RNA viruses to the total viroplankton. A total of 2481 putative RNA viral contigs (>500 bp) and 107 larger bona fide RNA viral genomes (>2.5 kb) were identified; 88 of them representing novel viruses belonging mostly to two clades: Yangshan assemblage (sister clade to the class Alsuviricetes) and Nodaviridae. These viruses were highly endemic and locally abundant, with little or no presence in other oceans since only ≈15% of them were found in at least one of the Tara sampling metatranscriptomes. Quantitative data indicated that the abundance of RNA viruses in the surface and deep chlorophyll maximum zone was within ≈106 VLP/mL representing a potential contribution of 5.2%-24.4% to the total viroplankton community (DNA and RNA viruses), with DNA viruses being the predominant members (≈107 VLP/mL). However, for the deep sample, the observed trend was the opposite, although as further discussed, several biases should be considered. Together these results contribute to our understanding of the diversity, abundance, and distribution of RNA viruses in the oceans and provide a basis for further investigation into their ecological roles and biogeography.

Estudio descriptivo de las infecciones asociadas con instrumentación de columna en un hospital terciario

IntroductionSpinal instrumentation-related infections (SIRI) are one of the main causes of post-surgical complication and comorbidity. Our objective was to describe the clinical and microbiological characteristics, treatment and prognosis of these infections. Material and methodsWe conducted a retrospective study in our institution (2011-2018) including adult patients undergoing spinal instrumentation who met the diagnostic criteria for confirmed infection. Superficial surgical wound and deep intraoperative samples were processed for microbiological culture. The medical and orthopaedic team was always the same. ResultsForty-one cases were diagnosed of which 39 patients (95.1%) presented early infection (<3 months after initial surgery) with symptoms in the first two weeks, mean CRP at diagnosis was 133mg/dl and 23% associated bacteremia. The remaining two patients (4.8%) were chronic infections (symptoms >3 months after surgery). The treatment of choice in early infections was the Debridement, Antibiotics and Implant Retention (DAIR) strategy without removal of the bone graft, which successfully resolved 84.2% of the infections. The main etiology was gram-positive (Staphylococcus aureus: 31.7%), followed by gram-negative and polymicrobial flora. Antibiotics were optimized according to cultures with a mean duration of 12 weeks. ConclusionsIn early infections, early diagnosis and DAIR strategy (with bone graft retention) demonstrated a healing rate higher than 80%.

CoffeeNet: A deep learning approach for coffee plant leaves diseases recognition

Coffee is regarded as the highest consumed drink around the globe and has accounted as a major source of income in the regions where it is cultivated. To meet the coffee marketplace's requirements around the globe, cultivators must boost and analyze its cultivation and quality. Several factors like environmental changes and plant diseases are the major hindrance to increasing the yield of coffee. The development in the field of computer vision has facilitated the earliest diagnostic of diseased plant samples, however, the incidence of various image distortions i.e., color, light, size, orientation changes, and similarity in the healthy and diseased portions of examined samples are the major challenges in the effective recognition of various coffee plant leaf infections. The proposed work is focused to overwhelm the mentioned limitations by proposing a novel and effective DL model called the CoffeeNet. Explicitly, an improved CenterNet approach is proposed by introducing spatial-channel attention strategy-based ResNet-50 model for the computation of deep and disease-specific sample characteristics which are then classified by the 1-step detector of the CenterNet framework. We investigated the localization and cataloging outcomes of the suggested method on the Arabica coffee leaf repository which contains the images captured in the more realistic and complicated environmental constraints. The CoffeeNet model acquires a classification accuracy number of 98.54%, along with an mAP of 0.97 that is presenting the usefulness of our technique in localizing and categorizing various sorts of coffee plant leaf disorders.

Factors Associated With the Development of a Confirmed Fracture-Related Infection.

To identify factors associated with the development of fracture-related infection (FRI) in patients who have undergone operative fixation of their fracture. Retrospective study. Academic medical center. Patients with peripheral extremity long-bone fractures and shoulder and pelvic girdle fractures who underwent operative repair. This included patients who had undergone fixation inside and outside of our institution. Chi-square, Fisher exact testing, analysis of variance (ANOVA), and t tests were used for analysis of data, as appropriate. Binomial logistic regression analysis was performed to determine risk factors of FRI. Demographics, fracture location, injury mechanism, open wound status, and wound complication type. One hundred ninety-three patients comprised the FRI cohort. Of those with confirmed signs of FRI, 36 (18.7%) had wound breakdown to bone or implant, 120 (62.1%) had a sinus track, and 170 patients (88.1%) had the confirmed presence of microorganisms in deep tissue samples obtained during operative intervention. Factors associated with development of an FRI were as follows: higher BMI; being of Black and Hispanic race; higher American Society of Anesthesiologists class; history of tobacco, alcohol, and drug use; and fractures of the lower extremity (foot, tibia, and femur/pelvis). The results of binary logistic regression demonstrated that treatment at an urban level I trauma center and drug use positively predicted the development of FRI. Our results demonstrate that higher BMI; being of Black or Hispanic race; higher American Society of Anesthesiologists; history of tobacco, alcohol, and drug use; and fractures of the foot, tibia, and femur/pelvis are all factors associated with development of FRI. Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.

점군 위치 정보 인코딩 활용을 통한 딥러닝 기반 포인트 클라우드 샘플링 성능 개선 연구

점군 위치 정보 인코딩 활용을 통한 딥러닝 기반 포인트 클라우드 샘플링 성능 개선 연구

In vivo Comparison of Positive Microbial Culture by Wound Irrigation Methods: Biportal Endoscopic versus Open Microscopic Transforaminal Lumbar Interbody Fusion.

This study aimed to evaluate the risk factors and prevalence of intraoperative contamination (IoC) through the microbial culture of superficial and deep samples obtained during surgery. Surgical site infection (SSI) in spinal surgery is a serious complication. The prevalence of IoC may differ based on surgical approach and technique, even in the setting of the same procedure. In this in-vivo study, microbial cultivation was performed with superficial (ligamentum flavum, LF) and deep (nucleus pulposus, NP) surgical specimens to evaluate IoC in 132 patients undergoing single-level transforaminal lumbar interbody fusion (TLIF). Biportal endoscopic (BE) TLIF was performed under continuous wound irrigation (group A, n=66), whereas open microscopic (OM) TLIF was performed under intermittent wound irrigation (group B, n=66). LF and NP specimens were homogenized, gram-stained, and cultured in aerobic and anaerobic media for 14 days. Microbial culture results and occurrence of SSI in the two groups were assessed. The Chi-square test and Fisher's exact test were used to determine significant differences among categorical variables. Logistic regression analysis was used to assess the influence of patient characteristics on the prevalence of positive microbial cultures. Of the 132 patients, 34 (25.8%) had positive microbial cultures, and positive culture required an incubation period of 72 h to 2 weeks in all these patients except for three. Overall positive culture was significantly higher in group B than in group A (P=0.029). The subgroups of LF- and NP-positive cultures were 18.18% (n=24) and 12.88% (n=17), respectively; the SSI was 0.76% (n=1). Group A had a significantly lower subgroup of NP-positive culture than group B (P=0.035). OM technique was an independent risk factor associatd with overall positive culture (P<0.05). The most common microorganism was Cutibacterium acnes (C. acnes). BE-TLIF with continuous wound irrigation showed significantly lower overall, and NP-positive cultures, than OM-TLIF with intermittent irrigation. The most common strain of positive culture was C. acnes. II.

Grain growth of natural and synthetic ice at 0 °C

Abstract. Grain growth can modify the microstructure of natural ice, including the grain size and crystallographic preferred orientation (CPO). To better understand grain-growth processes and kinetics, we compared microstructural data from synthetic and natural ice samples of similar starting grain sizes that were annealed at the solidus temperature (0 ∘C) for durations of a few hours to 33 d. The synthetic ice has a homogeneous initial microstructure characterized by polygonal grains, little intragranular distortion, few bubbles, and a near-random CPO. The natural ice samples were subsampled from ice cores acquired from the Priestley Glacier, Antarctica. This natural ice has a heterogeneous microstructure characterized by a considerable number of air bubbles, widespread intragranular distortion, and a CPO. During annealing, the average grain size of the natural ice barely changes, whereas the average grain size of the synthetic ice gradually increases. These observations demonstrate that grain growth in natural ice can be much slower than in synthetic ice and therefore that the grain-growth law derived from synthetic ice cannot be directly applied to estimate the grain-size evolution in natural ice with a different microstructure. The microstructure of natural ice is characterized by many bubbles that pin grain boundaries. Previous studies suggest that bubble pinning provides a resisting force that reduces the effective driving force of grain-boundary migration and is therefore linked to the inhibition of grain growth observed in natural ice. As annealing progresses, the number density (number per unit area) of bubbles on grain boundaries in the natural ice decreases, whilst the number density of bubbles in the grain interiors increases. This observation indicates that some grain boundaries sweep through bubbles, which should weaken the pinning effect and thus reduce the resisting force for grain-boundary migration. Some of the Priestley ice grains become abnormally large during annealing. We speculate that the contrast of dislocation density amongst neighbouring grains, which favours the selected growth of grains with low dislocation densities, and bubble pinning, which inhibits grain growth, are tightly associated with abnormal grain growth. The upper 10 m of the Priestley ice core has a weaker CPO and better-developed second maximum than deeper samples. The similarity of this difference to the changes observed in annealing experiments suggests that abnormal grain growth may have occurred in the upper 10 m of the Priestley Glacier during summer warming.

Intestinal Villi-Inspired Mathematically Base-Layer Engineered Microneedles (IMBEMs) for Effective Molecular Exchange during Biomarker Enrichment and Drug Deposition in Diversified Mucosa.

The mucosa-interfacing systems based on bioinspired engineering design for sampling/drug delivery have manifested crucial potential for the monitoring of infectious diseases and the treatment of mucosa-related diseases. However, their efficiency and validity are severely restricted by limited contact area for molecular transfer and dissatisfactory capture/detachment capability. Herein, inspired by the multilayer villus structure of the small intestine that enables high nutrient absorption, a trigonometric function-based periodic pattern was fabricated and integrated on the base layer of the microneedle patch, exhibiting a desirable synergistic effect with needle tips for deep sample enrichment and promising molecular transfer, significantly improving the device-mucosa bidirectional interaction. Moreover, mathematical modeling and finite element analysis were adopted to visualize and quantify the microcosmic molecular transmission process, guiding parameter optimization in actual situation. Encouragingly, these intestinal villi-inspired mathematically base-layer engineered microneedles (IMBEMs) have demonstrated distinguished applicability among mucosa tissue with varying surface curvatures, tissue toughness, and local environments, and simultaneously, have gained favorable support from healthy volunteers receiving preliminary test of IMBEMs patches. Overall, validated by numerous in vitro and in vivo tests, the IMBEMs were confirmed to act as a promising candidate to facilitate mucosa-based sampling and topical drug delivery, indicating highly clinical translation potential.

An effective technique for detecting minority attacks in NIDS using deep learning and sampling approach

Anomaly-based intrusion detection system have been consistently used in business organizations and military to detect a breach in network by identifying any activity that deviates from the baseline pattern. In this paper, we propose an effective intrusion detection technique to identify and predict the minority attacks with three layers. Here, the first layer utilizes a Weighted Deep Neural Network (WDNN) for identifying the suspicious traffic samples in network and it is passed to the second layer. Layer 2 classifies the traffic samples as normal or majority and minority attacks using Convolutional Neural Network (CNN) and Long-Short Term Memory (LSTM). Any traffic sample classified as minority attack is sent to Layer 3 that utilizes XGBoost algorithm. Layer 3 classifies the samples into their respective minority attack classes. To boost the detection rate of minority attacks, system employs a One-Sided Selection under-sampling algorithm to remove noisy samples from the majority attack classes. An Adaptive Synthetic (ADASYN) oversampling algorithm generates synthetic samples of minority attack classes. To evaluate the system, the datasets namely NSL KDD, CICIDS-2017 and CIDDS 001 dataset are used. The system attained an overall accuracy of 97.94% on NSL KDD dataset, 98.3% on CICIDS-2017 dataset and 97.9% on CIDDS 001 dataset.

Seasonal and spatial variability in bio-optical properties of the Persian Gulf: Implications for ocean color remote sensing

In this study, the absorption coefficients of phytoplankton community (aPh(λ)), Colored Dissolved Organic Matter (CDOM) [aCDOM(λ)], and Non-Algal Particles (NAP) [aNap(λ)] were analyzed in the northern Persian Gulf, a Mediterranean-type semi-enclosed marginal sea, through seven cruise expeditions conducted from May 2018 to February 2022. The highest and lowest aPh(443) values were observed in the shallow area (depth ≤20 m) during the wet season (November–April) (0.043 ± 0.004 m-1) and within the deep area (depth >20 m) through the dry season (July–September) (0.024 ± 0.003 m-1), respectively. The chlorophyll-specific absorption (a*Ph(λ)) values increased from the deep to shallow areas as well as the wet to dry seasons. A significant difference was further detected in a*Ph(λ) magnitude at the blue (443 nm) and red (675 nm) wavelengths, with the mean values of 0.0124/0.0223 m2 mg−1 and 0.0069/0.0135 m2 mg−1 in the wet/dry seasons, respectively. The mean values of aCDOM(443) in the shallow area during the wet season (0.10 ± 0.001 m-1) were significantly larger than those in the deep area during the dry season (0.06 ± 0.002 m-1). Similarly, aNAP(443) decreased from the shallow (0.056 ± 0.021 m-1) to deep areas (0.019 ± 0.011 m-1), and from the wet (0.035 ± 0.020 m-1) to dry (0.029 ± 0.015 m-1) seasons. The statistical analysis revealed that the differences in aCDOM(λ) and aNAP(λ) spectral distribution were not statistically significant for the duration of the wet and dry seasons, whereas significant differences were observed in the shallow and deep samples. During both wet and dry seasons, the magnitude of the blue-green absorption spectra was dominated by CDOM, but the red bands were associated with phytoplankton dominated samples. The ratio of aCDOM(λ)/a*Ph(λ) was utilized as an index for uncertainty assessment to estimate Chl-a concentrations by the National Aeronautics and Space Administration (NASA) standard Chl-a algorithms for the Moderate Resolution Imaging Spectrometer (MODIS; OC3M) and Sentinel-3 Ocean and Land Color Instrument (OLCI) (OLCI; OC4v6). The results additionally demonstrated that the satellite-derived Chl-a was overestimated due to higher aCDOM(λ) and underestimated owing to lower a*Ph(λ) at the low (<0.5 mg m−3) and high (>3 mg m−3) concentrations of Chl-a, respectively. The study findings established that the red bands should be considered for tuning or developing new Chl-a algorithms in the Persian Gulf.

Deep ensemble-based hard sample mining for food recognition

Deep neural networks represent a compelling technique to tackle complex real-world problems, but are over-parameterized and often suffer from over- or under-confident estimates. Deep ensembles have shown better parameter estimations and often provide reliable uncertainty estimates that contribute to the robustness of the results. In this work, we propose a new metric to identify samples that are hard to classify. Our metric is defined as coincidence score for deep ensembles which measures the agreement of its individual models. The main hypothesis we rely on is that deep learning algorithms learn the low-loss samples better compared to large-loss samples. In order to compensate for this, we use controlled over-sampling on the identified ”hard” samples using proper data augmentation schemes to enable the models to learn those samples better. We validate the proposed metric using two public food datasets on different backbone architectures and show the improvements compared to the conventional deep neural network training using different performance metrics.

The occurrence, spatial distribution, and well-depth dependence of PFASs in groundwater from a reclaimed water irrigation area

With the growing development of modern agriculture and industry, groundwater is facing more and more complex contaminants. One such contaminant is per- and polyfluoroalkyl substances (PFASs), which pose a potential risk to human health, particularly for those who rely on groundwater as their primary source of drinking water. In this study, we conducted a comprehensive investigation on the occurrence, spatial distribution, and source apportionment of PFASs in shallow (<60 m) and deep (>80 m) groundwater samples from a reclaimed water irrigation area in Beijing's suburbs. Our results showed that the average total PFAS concentration (∑10PFAS) for all samples was 10.55 ± 7.77 ng/L, ranging from 1.05 to 34.28 ng/L. The dominant congeners were PFBA, PFOA, and PFBS. No significant linear relationship was observed between PFAS concentrations and the well depth. However, the averaged ΣPFASs in groundwater were highest in the uppermost layer and declined sharply to a few ng/L in the deep aquifer below 80 m. PFASs showed elevated concentration in shallow aquifers in 9 out of 11 paired wells, indicating an overall descending trend of PFASs with increasing aquifer depth. The spatial distribution of PFASs was highly heterogeneous and showed different patterns in shallow and deep groundwater, which may be related to the complicated attenuation behavior of PFAS compounds when they transport and diffuse through overlapping aquifer layers. The influence of the landfill on groundwater PFASs was most pronounced within a 5 km radius. Source apportionment results indicated that reclaimed water irrigation is the main non-point source of PFASs in shallow groundwater. In contrast, deep groundwater is primarily subject to point sources and lateral recharge flow. This investigation of PFASs in shallow and deep wells provides a foundation for further exploration of PFASs transportation and risk prevention in regions where groundwater is a major water resource for domestic and industrial development.