Local Data Research Articles

Evaluation of the Level of Electrolytes in Children with Steroid Sensitive or Steroid Resistant Nephrotic Syndrome

Background: Childhood idiopathic nephrotic syndrome is one of the most common conditions pediatric nephrologists encounter globally. Nephrotic syndrome is characterized by proteinuria, hyperlipidemia, and edema. The degree and duration of proteinuria have an impact on serum electrolyte levels. However, local data is limited. Objectives: To assess and compare the degree of electrolyte imbalance and its relationship to kidney function indicators during relapse and remission in children with idiopathic nephrotic syndrome. Methods: In this case-control study, blood samples were collected from 80 Iraqi children with an age range of (2-14) years. They were divided into three groups: Group I (20 individuals with steroid-sensitive nephrotic syndrome (SSNS)), Group II (20 individuals with steroid-resistant nephrotic syndrome (SRSN)), and Group III (40 healthy individuals as the control group). Serum electrolyte levels (Na, Ca, Cl, and K) were measured by an ion-selective electrode (9180 electrolyte analyzer). Blood creatinine and urea were measured by a Cobas c311 autoanalyzer during the relapse and remission phase. The patients were clients of the pediatric nephrology consultation center at the Children's Teaching Hospital / Baghdad Medical City and Al-Batoul Teaching Hospital for Women and Children from 15 February to 20 August 2022. The controls were healthy children whose medical history was reviewed to eliminate a history of kidney disease and underwent a comprehensive physical examination. Controls were recruited from a network of family, friends, relatives, and the National Autism Center/ Child Protection Teaching Hospital affiliated with Medical City/Baghdad. Results: Serum Calcium levels showed a clear decrease in all SSNS and SRSN patients compared to the control group. The levels of Sodium and Chloride were significantly lower than the control group during the relapse phase. The results of the relapse phase of SRNS patients indicated higher serum potassium concentration compared with the control group and the SSNS patient group, with a statistically significant difference). Conclusion: All children with idiopathic nephrotic syndrome had hypocalcemia in the relapse and remission phase. SRNS cases had hyperkalemia, Sodium, and chloride fluctuated between low levels during the relapse phase and normal levels during the remission phase.

FAM: Adaptive federated meta-learning for MRI data

Federated learning enables multiple clients to collaborate to train a model without sharing data. Clients with insufficient data or data diversity participate in federated learning to learn a model with superior performance. MRI data suffers from inadequate data and different data distribution due to differences in MRI scanners and client characteristics. Also, privacy concerns preclude data sharing. In this work, we propose a novel adaptive federated meta-learning (FAM) mechanism for collaboratively learning a single global model, which is personalized locally on individual clients. The learnt sparse global model captures the common features in the MRI data across clients. This model is grown on each client to learn a personalized model by capturing additional client-specific parameters from local data. Experimental results on multiple data sets show that the personalization process at each client quickly converges using a limited number of epochs. The personalized client models outperformed the locally trained models, demonstrating the efficacy of the FAM mechanism. Additionally, the FAM-based sparse global model has fewer parameters that require less communication overhead during federated learning. This makes the model viable for networks with limited resources.

FedDA: Resource-adaptive federated learning with dual-alignment aggregation optimization for heterogeneous edge devices

Federated learning (FL) is an emerging distributed learning paradigm that allows multiple clients to collaborate on training a global model without sharing their local data. However, in practical heterogeneous edge device scenarios, FL faces the challenges of system heterogeneity and data heterogeneity, which leads to unfair participation and degraded global model performance. In this paper, we introduce FedDA, a resource-adaptive FL framework, which adapts to the client’s computing resources by assigning heterogeneous models of different sizes. To improve the performance of heterogeneous model aggregation and adjust to non-independent and identically distributed (non-i.i.d.) data, we propose a dual-alignment aggregation optimization method, i.e., parameter feature space alignment and output space alignment. Specifically, FedDA exploits the permutation symmetry property of weight space to permutate the model parameter positions via an adaptive layer-wise matching method, thereby aligning models with significant deviations in parameter feature space. FedDA mitigates the imbalance in parameter quantity between smaller and larger models through parameter expansion. Additionally, FedDA maps client labels into a uniform embedding space through output space alignment, thus reducing model parameter deviations due to non-i.i.d. data without additional client-side computing overhead. We evaluate the performance of FedDA on benchmark datasets, including FashionMNIST, CIFAR10, CIFAR100 and AGNews. Experimental results demonstrate that FedDA achieves up to 8.71% improvement in model accuracy compared to baseline methods, highlighting its effectiveness in addressing the challenges of heterogeneity.

Multi-temporal assessment of a wildfire chronosequence by remote sensing

The study aimed to develop a methodological framework to identify forest ecosystems affected by wildfires and evaluate their recovery chronologically. To do this remote sensing analysis, sites with burn scars were selected based on various criteria (fire severity, affected area, vegetation and soil type, slope, aspect, and one-time occurrence of wildfire in the last 23 years). Spectral vegetation indices (VIs) from satellite imagery were used to estimate burn severity and vegetation cover changes. Images of surface reflectance were obtained from the collection of Landsat 5 ETM, Landsat 7 ETM+, and Landsat 8 OLI/TIRS, available and processed on the Google Earth Engine Platform (GEE). Indices VIs (i) the normalized difference vegetation index (NDVI), (ii) the normalized burn ratio (NBR), and (iii) the differenced normalized burn ratio (dNBR) were calculated to classify burn severity. The one-time occurrence selection was performed using the LandTrendr algorithm to monitor changes in land cover and burned areas. To validate the selection, the chosen sites within the chronosequence were clustered on 4 seasons of soil properties and litter accumulation recovery. Our result can guide methodological comparisons and forest management practices on large surfaces by comparing parches of different time-affected ecosystems. Validation sites of the cluster chronosequence shows consistent recovery of soil properties as soil carbon, bulk density and litter accumulation through the studied years•The study developed a framework to identify wildfire-affected forest ecosystems and evaluate their recovery using remote sensing and local data.•Vegetation indices (NDVI, NBR, dNBR) from Landsat satellite imagery processed on the Google Earth Engine were used to assess burn severity and vegetation changes over time.•Selected sites were validated using the LandTrendr algorithm and monitored for seasonal changes in soil properties and litter accumulation.

An efficient hyperspectral image classification method using retentive network

In recent computer vision tasks, the vision transformer (ViT) has demonstrated competitive ability. However, ViT still has problems: the computational complexity of the self-attention layer leads to expensive and slow interference, and processing all tokens for high-resolution images may slow down due to the layer’s quadratic complexity. Recently, a retentive network with excellent performance, training parallelism, and an inexpensive inference cost was proposed. For hyperspectral image (HSI) classification, this paper proposesa retention-based network model called the HSI retentive network (HSIRN). The proposed model allows memory usage independent of the token’s sequence, facilitating the efficient processing of high-resolution images with low inference and computational costs. Although the retention encoder can extract global data, it pays limited attention to local data. A powerful tool for extracting local information is a convolutional neural network (CNN). The proposed HSIRN model uses a specific CNN-based block to extract local spectral-spatial information. To maintain degradation between successive vertical and horizontal positions with the depth dimension of the HSI, we propose a three-dimensional retention mechanism for the three-dimensional HSI dataset in the retention encoder. By efficiently using both local and global spectral-spatial information, the proposed method offers a potent tool for HSI classification. We evaluated the classification performance of the proposed HSIRN approach on four datasets through comprehensive examinations, and the results demonstrated its superiority over state-of-the-art methods. At https://github.com/RajatArya22/HSIRN, the source code will be available to the public.

Identifying the Original Source of Megalithic Stones in Tambunan Using Spectral Signature Technique

Megalithic stones are a remnant of the ancient people. An intriguing megalithic stone legacy ought to emphasize the context in which it fits. Consequently, it is necessary to ascertain each megalithic stone’s historical context. This is significant for the archaeologist to obtain true and valid information. In Tambunan, there are numerous myths about megalithic stones. As a result, it is impossible to verify the story’s validity. These days, remote sensing has emerged as a crucial technical instrument for archaeologists to gather spatial data in both broad and localized locations. Prior research has demonstrated the efficacy of remote sensing in identifying archaeological site remnants. Nevertheless, there have never been any investigations utilizing remote sensing in relation to small-sized megalithic stones. Thus, this study employs spectral signature techniques and remote sensing to identify the original source of the megalithic stones discovered in Tambunan. This study carries out the spectral signature of megalithic stones in Tambunan using Sentinel-2A satellite image data with a 5-meter resolution using the SNAP application. Before that, the coordinates and measurements for each megalithic stone were obtained from the ground. The spectral signature of all the megalithic stones was then compared with the spectral signature of the stones near the river or near the hill. The study’s findings demonstrate that the spectral signature of the megalithic stones in question is identical to the stone found in the river near its location. As a result, this study was successful in providing scientific evidence that the river is the source of the megalithic stones in Tambunan, despite the stones’ considerable distance from their site.

Getting mammograms shouldn’t be so bumpy: Improving mammogram workflow in VA community care.

345 Background: The interplay between VA-based care and Community Care (CC) has long been important for continuity of patient care, though transitioning care between the two is often fraught with problems. Mammograms ordered at VA San Diego (VASD) are often deferred to Community Care; while this process should be smooth and seamless, it is instead challenging to navigate, with inconsistent rates of completion and results often not relayed to patients leading to delayed/missed diagnoses. We applied extensive QI methodology towards overhauling the VASD Community Care Mammogram workflow to improve rates of CC mammogram completion and results released to patients in a timely manner. Methods: We evaluated the CC mammogram workflow and implemented interventions at VASD from 01/2022-04/2023. An A3 was used for tracking process improvement. Using Lean Six Sigma process improvement strategies, we identified several problematic areas within the CC Mammogram workflow that are prone to errors and can lead to patient safety issues. Weekly meetings/interviews were set up to identify pain points with the primary users to allow for continuous process evaluation and improvement. Primary outcome was the frequency with which a CC mammogram order had mammogram completed and result released to patient within the expected timeframe by VISN 22 metrics (BIRADS 0,3,4,5,6 <7 days and BIRADS 1,2 <14 days to notification). Results: Process mapping of the VASD CC mammogram process revealed significant waste and failure modes. Notably among these was the splintered tracking systems used across multiple users and lack of automated follow-up reminders. Multiple PDSAs involved standardizing workflow by developing an SOP that focused on standardization of tracking orders/results, creating automated task reminders, while also adding backup systems and double-checks. To centralize and standardize the patient tracking process, we repurposed the Coordinated Care Tracking System (CCTS), a VA-based tool that is used for tracking patients in other departments and creating reminders for follow up, while storing patient data in a centralized location. We also redesigned and streamlined the mammogram ordering menu within CPRS. Implementation of the above interventions increased monthly rates of mammograms completed and reported to patients in a timely manner from a median of 67.5% to 100% in all BIRADS groups by 10/2022, and sustained at 100% through 04/2023. Conclusions: Standardization of workflow and incorporation of CCTS software into the CC Mammogram process led to an appreciable improvement in rates of CC mammogram scheduling and timely result notifications. Improved tracking of CC mammogram patients has the potential to prevent unnecessary patient harm. It is worth considering whether CCTS and a similar standardized process should be incorporated into other CC programs, as these issues are not unique to VASD mammograms.

Peat profile database from peatlands in Canada.

Peatlands cover approximately 12% of the Canadian landscape and play an important role in the carbon cycle through their centennial- to millennial-scale storage of carbon under waterlogged and anoxic conditions. In recognizing the potential of these ecosystems as natural climate solutions and therefore the need to include them in national greenhouse gas inventories, the Canadian Model for Peatlands module (CaMP v. 2.0) was developed by the Canadian Forest Service. Model parameterization included compiling peat profiles across Canada to calibrate peat decomposition rates from different peatland types, to define typical bulk density profiles, and to describe the hydrological (i.e., water table) response of peatlands to climatic changes. A total of 1217 sites were included in the dataset from published and unpublished sources. The CORESITES table contains site location and summary data for each profile, as well as an estimate of total carbon mass per unit area (in megagrams of C per hectare). Total carbon mass per unit area at each location was calculated using bulk density and carbon content through each profile. The PROFILES table contains data for depth (in centimeters), bulk density (in grams per cubic meter), ash and carbon content (in percentage), and material descriptions for contiguous samples through each peat profile. Data gaps for bulk density and C content were filled using interpolation, regression trees, and assigned values based on material description and/or soil classification to allow for the estimation of total carbon mass per unit area. A subset of the sites (N = 374) also have pH and pore water trace-elemental geochemistry data and are found in the WATER table. The REFERENCES table contains the full citation of each source of the data and is linked to each core location through the SOURCEDATA table. The LOOKUP table defines codes in the database that required more space that what was sufficient in the metadata tables. The data can be accessed on Open Government Canada and will be useful for future work on carbon stock mapping and ecosystem modeling. All metadata and data are provided © Her Majesty the Queen in Right of Canada, 2023 and information contained in this publication may be reproduced for personal or public noncommercial purposes with attribution, whereas commercial reproduction and distribution are prohibited except with written permission from NRCan; complete details are noted in the Supporting Information file Metadata S1 (see Class III.B.3: Copyright restrictions).

EcD-Net: Encoder-Corollary Atrous Spatial Pyramid Pooling-Decoder Network for Automated Pancreas Segmentation of 2D CT Images

Automatic pancreas segmentation of CT scans enables physicians to identify and monitor the abnormalities in the pancreas. This facilitates intraoperative assistance, surgical planning, prognosis, and diagnosis. Nonetheless, the size and location of the pancreas in the CT image input data present a significant challenge for automatic segmentation, and the intricacy of the background region confounds deep segmentation networks. To solve this difficulty, we propose the Encoder-Corollary Atrous Spatial Pyramid Pooling-Decoder Network (EcD-Net) for locating and segmenting the pancreas. This two-tiered method begins with a coarse segmentation stage for locating the pancreas within the overall CT image. Using the detected image from the first tier, a fine segmentation network based on U-Net is applied to segment the target organ (pancreas). A novel Saturated Multi-view Dense Module (SMD- Module) is presented to enhance information gradient flow for the stability of the training process and easier convergence at the fine stage. A novel Corollary Atrous Spatial Pyramid Pooling Module (CASPP-Module) is proposed to simultaneously capture low-level details and high-level global context information to enhance the pancreas segmentation accuracy, extract global and local spatial data, and detect the pancreas at varied scales. On the publicly available National Institute of Health (NIH) pancreas dataset, our proposed EcD-Net surpasses previous state-of-the-art methods with a mean Dice Similarity Coefficient (DSC) of 88.84%, mean Precision (PRE), Recall (REC), and Intersection over Union (IoU) of 91.85%, 88.21%, and 89.94% respectively, and it also records the lowest standard deviations showing the robustness and steadiness of our proposed method.

Practical and Robust Federated Learning With Highly Scalable Regression Training.

Privacy-preserving federated learning, as one of the privacy-preserving computation techniques, is a promising distributed and privacy-preserving machine learning (ML) approach for Internet of Medical Things (IoMT), due to its ability to train a regression model without collecting raw data of data owners (DOs). However, traditional interactive federated regression training (IFRT) schemes rely on multiple rounds of communication to train a global model and are still under various privacy and security threats. To overcome these problems, several noninteractive federated regression training (NFRT) schemes have been proposed and applied in a variety of scenarios. However, there are still several challenges: 1) how to protect the privacy of DOs' local dataset; 2) how to realize highly scalable regression training without linear dependence on sample dimension; 3) how to tolerate DOs' dropout; and 4) how to enable DOs to verify the correctness of aggregated results returned from the cloud service provider (CSP). In this article, we propose two practical noninteractive federated learning schemes with privacy-preserving for IoMT, named homomorphic encryption based NFRT (HE-NFRT) and double-masking protocol based NFRT (Mask-NFRT), respectively, which are based on a comprehensive consideration of NFRT, privacy concerns, high-efficiency, robustness, and verification mechanism. The security analyses display that our proposed schemes are able to protect the privacy of DOs' local training data, resist collusion attack, and support strong verification to each DO. The performance evaluation results demonstrate that our proposed HE-NFRT scheme is desirable for a high-dimensional and high-security IoMT application while Mask-NFRT scheme is desirable for a high-dimensional and large-scale IoMT application.

Quantum data compression under localized features

In this paper, we first introduce the local quantum information processing task by compressing the density operator based on local quantum Bernoulli noises. Then, the local quantum compression theorem is given, that is, the local quantum entropy is the minimum achievable rate of local compression. Finally, we prove the theorem with the proof of the direct coding theorem and the inverse theorem. The direct coding theorem shows that a scheme with such a local compression rate exists, and that this local compression rate converges to the local quantum entropy. The inverse theorem shows that the compression scheme with the rate below the local entropy is unachievable. With the continuous development of quantum technology, local quantum data compression technology will have broad application prospects and development space.

Validating the Vegetable Variety Navigator Decision-support Tool Using Recent On-farm Variety Trial Data

Local variety trial data are necessary for informing growers how a specific variety might perform on their farm but there is a growing deficiency in these data, particularly for specialty crops. To address this issue, an online decision-support tool named the Vegetable Variety Navigator (VVN) was developed in 2020 to compile, analyze, and visually communicate publicly available broccoli (Brassica oleracea var. italica), cucumber (Cucumis sativus), and sweet pepper (Capsicum annuum) variety trial data. To validate the accuracy and predictive potential of the VVN, we conducted 16 on-farm variety trials for broccoli, cucumber, and sweet pepper between 2020 and 2022. Yield of each variety in a trial was compared with the mean of all other varieties in the same trial to calculate a mean relative yield (MRY). The difference between observed relative yield in the field and predicted relative yield from the VVN (ΔMRY) was used to assess the accuracy of three different VVN prediction strategies. Strategies included using data from the following: 1) the single geographically nearest trial, 2) the mean of the three geographically nearest trials, or 3) the mean of all available trial data regardless of location. Compared with random predictions of MRY for each variety (from within a normal distribution of MRY values in the VVN database), the VVN predictions reduced ΔMRY and improved the accuracy of relative yield predictions across varieties by up to 21% in broccoli, 51% in cucumber (depending on strategy), and 51% in sweet pepper. Results confirm the value of the VVN as a decision-support tool for growers facing an ever-increasing number of crop variety options with less variety trial data. Future research and development are needed to improve the accuracy of the VVN by accounting for possible effects of crop types, management, and location on relative yield and quality.

GOLM1 promotes prostate cancer progression via interaction with PSMD1 and enhancing AR-driven transcriptional activation.

Aberrant transcriptional activation of the androgen receptor (AR) is a predominant cause of prostate cancer (PCa), including both in the initial and androgen-independent stages. Our study highlights Golgi membrane protein 1 (GOLM1) as a key regulator of AR-driven transcriptional activity in PCa progression. Utilizing local clinical data and TCGA data, we have established a robust association between GOLM1 and AR target genes, and further demonstrated that GOLM1 can enhance the expression of AR target genes. We discovered that GOLM1 interacts with PSMD1, a component of the 19S regulatory complex in the 26S proteasome, using mass spectrometry and Co-IP analysis. It is well known that ubiquitin-proteasome plays a vital role in AR expression and transcriptional regulation. Our findings demonstrate that GOLM1 enhances ubiquitin proteasome activity by binding to PSMD1, thereby facilitating AR-driven transcriptional activity and PCa progression. These results indicate that GOLM1 and its associated proteins may become potential therapeutic targets for PCa characterized by dysregulated AR-driven transcriptional activation.

Predictive factors associated with Bile culture positivity And phenotypiCal antIbiogram resistance patterns in patients taken to LaparOscopic cholecystectomy (BACILO): protocol for a prospective observational cohort study and development of a prognostic prediction model.

Bile fluid is aseptic under normal conditions; however, in the presence of cholecystitis, its susceptibility to bacterial colonisation varies, with reported rates of 20%-70% of cases. This process is referred to as bactibilia and/or bacteriobilia and can be considered a secondary complication of biliary stasis and cholecystitis in general. In the management of acute cholecystitis, the antibiotic regimen should be prescribed based on the presumed pathogens involved, taking into consideration the risk factors for resistance patterns according to demographics and local exposure. The aim of this study is to determine the predictive factors for bile culture positivity and antibiotic resistance in patients who underwent laparoscopic cholecystectomy in the Méderi Hospital Network. We hope to develop a predictive model that allows us to better guide antibiotic therapy. This is a prospective observational cohort study with prognostic prediction model. Patients who will undergo laparoscopic cholecystectomy and have bile cultures taken in the Méderi Hospital Network during the study period will be included. The dependent variables will be positive bile culture and antibiotic resistance, and the predictive variables will be age, presence of diabetes, diagnosis of choledocholithiasis, diagnosis of cholecystitis and severity of cholecystitis according to the Tokyo criteria. The minimum sample size has been calculated at 703 patients. Follow-up will continue until a control appointment 15 days after the procedure. The primary outcomes are bile culture positivity and phenotypical antibiogram resistance. For each outcome, a multivariate logistic regression will be performed using frequentist and Bayesian prediction techniques. This study was approved by the Méderi network research department committee (CIMED) and by Universidad del Rosario's Research Ethics Committee (CEI-UR; DVO005 2555-CV1837). Written informed consent is required for participation. The results will be disseminated through the submission of an academic article to a high-impact scientific journal, presentations at academic conferences, and sharing with our institution's faculty to inform antimicrobial therapy management based on local epidemiological data. NCT06314399.

Comparative outcomes of selective laser trabeculoplasty delivered by optometrists compared with ophthalmologists: a UK-based multicentre observational study

BackgroundSelective laser trabeculoplasty (SLT), a National Institute for Care and Health Excellence recommended first-line treatment for open-angle glaucoma and ocular hypertension, is increasingly delivered by optometrists. This retrospective multicentre observational...

Park thermal comfort and cooling mechanisms in present and future climate scenarios

Extreme heat is Australia’s most perilous natural hazard, and increasing urban temperatures due to climate change are a growing concern. Consequently, there is growing interest in developing nature-based solutions (i.e., greenery and vegetated surfaces) to cool urban areas. Appropriately designed urban parks are anticipated to be crucial for maintaining thermal comfort as temperatures rise. The two main diurnal cooling mechanisms of urban parks are shade provision and vegetation transpiration. However, limited studies have examined the cooling performance of vegetation through transpiration, especially in the southern hemisphere. This study addresses this gap by examining the microclimatic conditions, cooling benefits, and thermal performance of a typical neighbourhood park in Perth, Western Australia, with a focus on the cooling performance of vegetation through shade and transpiration. Present and future microclimates were modelled and simulated for average and hottest summer days based on 25 years of local weather data and projections for 2090 under the Representative Concentration Pathway (RCP) 8.5 scenario. The findings reveal that trees provided diurnal cooling benefits for park users by lowering the Universal Thermal Comfort Index (UTCI) by up to 17°C, with this benefit persisting in projected 2090 conditions. This cooling benefit was predominantly achieved through shade provision, with marginal contributions from transpiration. Additionally, on hot days, as leaf temperature exceeded 30°C, increased stomatal resistance led to reduced transpiration. Therefore, more attention must be paid to transpiration cooling limits due to stomatal closure during hot hours to improve cooling performance in park design. Moreover, comparing different plant species’ behaviour and adaptability on hot days is crucial, especially in future climatic conditions.

HSMix: Hard and soft mixing data augmentation for medical image segmentation

Due to the high cost of annotation or the rarity of some diseases, medical image segmentation is often limited by data scarcity and the resulting overfitting problem. Self-supervised learning and semi-supervised learning can mitigate the data scarcity challenge to some extent. However, both of these paradigms are complex and require either hand-crafted pretexts or well-defined pseudo-labels. In contrast, data augmentation represents a relatively simple and straightforward approach to addressing data scarcity issues. It has led to significant improvements in image recognition tasks. However, the effectiveness of local image editing augmentation techniques in the context of segmentation has been less explored. Additionally, traditional data augmentation methods for local image editing augmentation methods generally utilize square regions, which cause a loss of contour information.We propose HSMix, a novel approach to local image editing data augmentation involving hard and soft mixing for medical semantic segmentation. In our approach, a hard-augmented image is created by combining homogeneous regions (superpixels) from two source images. A soft mixing method further adjusts the brightness of these composed regions with brightness mixing based on locally aggregated pixel-wise saliency coefficients. The ground-truth segmentation masks of the two source images undergo the same mixing operations to generate the associated masks for the augmented images.Our method fully exploits both the prior contour and saliency information, thus preserving local semantic information in the augmented images while enriching the augmentation space with more diversity. Our method is a plug-and-play solution that is model agnostic and applicable to a range of medical imaging modalities. Extensive experimental evidence has demonstrated its effectiveness in a variety of medical segmentation tasks. The source code is available in https://github.com/DanielaPlusPlus/HSMix.

Photometric redshifts probability density estimation from recurrent neural networks in the DECam local volume exploration survey data release 2

Photometric wide-field surveys are imaging the sky in unprecedented detail. These surveys face a significant challenge in efficiently estimating galactic photometric redshifts while accurately quantifying associated uncertainties. In this work, we address this challenge by exploring the estimation of Probability Density Functions (PDFs) for the photometric redshifts of galaxies across a vast area of 17,000 square degrees, encompassing objects with a median 5σ point-source depth of g= 24.3, r=23.9, i= 23.5, and z= 22.8 mag. Our approach uses deep learning, specifically integrating a Recurrent Neural Network architecture with a Mixture Density Network, to leverage magnitudes and colors as input features for constructing photometric redshift PDFs across the whole DECam Local Volume Exploration (DELVE) survey sky footprint. Subsequently, we rigorously evaluate the reliability and robustness of our estimation methodology, gauging its performance against other well-established machine learning methods to ensure the quality of our redshift estimations. Our best results constrain photometric redshifts with the bias of −0.0013, a scatter of 0.0293, and an outlier fraction of 5.1%. These point estimates are accompanied by well-calibrated PDFs evaluated using diagnostic tools such as Probability Integral Transform and Odds distribution. We also address the problem of the accessibility of PDFs in terms of disk space storage and the time demand required to generate their corresponding parameters.We present a novel Autoencoder model that reduces the size of PDF parameter arrays to one-sixth of their original length, significantly decreasing the time required for PDF generation to one-eighth of the time needed when generating PDFs directly from the magnitudes.

Exploring urban and agricultural land use planning

Socio-economic development has accelerated urban expansion, leading to the encroachment of urban space into agricultural space land and seriously undermining the stability of ecosystems. Scientifically optimizing the layout of urban-agricultural space can guarantee China's food security, promote urban development and maintain ecological stability. In this study, we focused on the rapidly growing city of Nankang. We used land, ecological and geographic location data to calculate land suitability, predicted future urban space through the GeoSOS-FLUS model, and finally optimized the agricultural-urban space conflict. The results show that: the areas of farmland with quality evaluation results for grades 1-4 were 7783.13 hm2, 9411.01 hm2, 8738.99 hm2 and 3032.73 hm2, and the areas suitable for construction, more suitable for construction, unsuitable for construction and prohibited from construction accounted for 12.19%, 16.23%, 41.34% and 30.24%, respectively. Combining the FLUS model with the land suitability evaluation results, we found that there is a conflict area of 1841.96 hm2 in the delineation scheme. Adjusting their layout according to the principle of conflict coordination, the optimized area of agricultural-urban space is 28429.34 hm2 and 8968.23 hm2 respectively, accounting for 16.32% and 5.15% of the study area. The result of spatial contradiction optimization satisfies the balance between production and residential space, promotes urban development, protects ecological security, and can provide scientific reference for regional urban ecological environment construction.

Trade-Offs Between Forage Availability, Accessibility, and Predation Risk on Winter Foraging Strategies of Wood Bison (Bison bison athabascae).

Optimal foraging theory (OFT) and the energy maximization hypothesis (EMH) have long been essential when examining wildlife habitat selection. At high latitudes and altitudes, animals in winter face greater limitations in availability and accessibility of forage. Here we explore the foraging behavior of wood bison (Bison bison athabascae) during winter within the Ronald Lake bison herd in northeastern Alberta, Canada, and examine the trade-offs they face due to limitations in forage abundance and availability (snow conditions), as well as the need to minimize predation risk. We used Global Positioning System (GPS) location data collected from 70 female wood bison to identify winter foraging sites and craters selected by bison to access forage beneath the snow. Within wetlands used by bison we selected 190 pairs of used (foraged) and random (available) sites to test eight a priori hypotheses explaining how bison traded-off between forage availability, accessibility, and minimizing predation risk. We found with matched-paired logistic regression that Carex atherodes was 1.21-times more likely to be selected per unit increase in ground cover, compared to 1.17-times per unit ground cover for C. aquatilis and C. utriculata. However, all Carex species showed an increase in selection when cover was > 50% cover within individual craters. While the importance of Carex was clear, forage site selection was still inversely related to snow depth. There is also a neutralizing combined effect of snow depth and Carex species ground cover which suggests that bison maximized their energy return by avoiding areas with deep snow (> 30 cm) that demanded intensive cratering, even when highly selected forage was accessible beneath. Avoidance of forage areas with deep snow demonstrates that wood bison employed a foraging strategy that considers both forage availability and environmental conditions, with snow depth being a limiting factor. We highlight the relationship between optimal foraging based on food availability and the trade-offs within an energy restrictive winter season, furthering the understanding of how large herbivores forage strategically to maximize energy intake in northern environments.