Search Results (983)

While systematic video analysis is widely used to understand anterior cruciate ligament (ACL) injury mechanisms and contexts, human observation reliability remains a source of concern. This study evaluates the intra- and inter-rater reliability of a systematic video-analysis checklist for assessing ACL injuries in elite men’s football. Twenty-five match-related injuries from the top six European leagues (2020–2024) were randomly selected. Independent observers assessed contextual and situational (sunny weather, match minute, playing phase, field location, injury side, dominant leg, and situational pattern), biomechanical (player contact and anatomical area of player contact), and neurocognitive (attentional inhibition and motor response inhibition) variables. Reliability was calculated using Cohen’s kappa ($\kappa$) and Intraclass Correlation Coefficients (ICC). Quantitative variables and macro-contextual factors, including injury side, playing phase, and situational pattern (0.810 < $\kappa$ < 1.000) revealed near-perfect to perfect agreement. Biomechanical details exhibited substantial agreement (0.601 < $\kappa$ < 0.784). Neurocognitive variables only reached moderate to substantial agreement (0.503 < $\kappa$ < 0.752), while visual speed estimations proved highly unreliable (−0.106 < $\kappa$ < 0.412). The checklist is a highly reliable tool for evaluating the contextual and situational patterns of ACL injuries, but visual speed estimation should be removed or replaced by objective tracking technologies. Full article

Accurate change detection in high-resolution remote sensing imagery is essential for urban planning, land-use monitoring, and disaster response. This study introduces CTA-Net, a Cross-Temporal Attention Network for binary change detection in bi-temporal optical imagery, designed to improve robustness against pseudo-changes caused by illumination variation, seasonal effects, and sensor noise. The proposed method employs a shared Siamese encoder with multi-scale Cross-Temporal Attention modules that derive spatial and channel attention from L2 feature differences, along with a lightweight confidence estimation head for per-pixel uncertainty modelling. A hybrid loss function combining confidence-weighted binary cross-entropy and focal loss is used to address class imbalance. Experiments on the LEVIR-CD dataset demonstrate that CTA-Net achieves an overall accuracy of 98.99%, an F1-score of 87.68%, an Intersection over Union of 78.06%, a Cohen’s kappa of 0.8715, and a Matthews Correlation Coefficient of 0.8721, with stable convergence and minimal overfitting. Qualitative and calibration analyses further indicate that the model produces interpretable attention maps and reliable probabilistic outputs. To evaluate cross-domain generalization, we conduct a transfer learning case study on multispectral Sentinel-2 agricultural imagery. The model is adapted to 11-channel input and fine-tuned on automatically generated change masks derived from NDVI-delta thresholding. Under this supervision protocol, CTA-Net achieves an F1-score of 95.18% and an IoU of 90.81% on a held-out test region, with balanced precision and recall. While these results demonstrate effective adaptation across sensor modality, spatial resolution, and semantic domain, the evaluation reflects agreement with the mask generation procedure rather than independently annotated ground truth. While CTA-Net shows strong performance and reasonable interpretability, its cross-domain evaluation is limited by the use of automatically generated labels. As a result, the reported transferability should be interpreted cautiously until validated on human-annotated datasets. Full article

Background and Objectives: Musculoskeletal ultrasound (MSUS) is increasingly used to assess structural changes in knee osteoarthritis (KOA), although measurement reproducibility may vary with examination protocol and operator experience. This pilot study aimed to evaluate intra- and inter-observer reliability for assessing cartilage thickness, osteophyte dimensions and meniscal displacement in supine and weight-bearing conditions in patients with KOA. Material and Methods: This prospective reliability study included patients with KOA stages 2–3 on the Kellgren–Lawrence (K-L) scale. Ultrasound (US) evaluation of cartilage thickness and osteophytes was performed in the supine position, while medial and lateral meniscal extrusion was additionally assessed in bipedal and single-leg conditions using a standardized protocol. Intra- and inter-reproducibility were evaluated through the intraclass correlation coefficient (ICC), agreement analysis and kappa statistics. The analysis was performed using Microsoft Excel (version 2019) and IBM SPSS version 26 software. Results: Both inter-reliability and intra-reliability were evaluated, with intra-rater agreement being constantly higher than inter-rater agreement. From the inter-rater reliability, ICC values ranged from 0.48 to 0.88 for quantitative ultrasound measurements, whereas osteophyte assessment showed Cohen’s k values ranging from 0.71 to 0.86, indicating substantial to almost perfect agreement. Both evaluators identified a higher number of knees with medial meniscal extrusion > 3 mm under weight-bearing conditions than in the supine position. Conclusions: Standardized US evaluation demonstrated at least good reproducibility for structural abnormalities in KOA. Medial and lateral meniscal displacement increased under loading conditions, highlighting the added value of dynamic weight-bearing assessment. However, the clinical significance of load-dependent meniscal extrusion requires further investigation. Full article

Background/Objectives: Metal artifacts from dental implants degrade CBCT image quality and may impair detection of peri-implant bone defects. Zirconia implants produce stronger artifacts than titanium due to its higher atomic number. This study evaluated the impact of a native MAR algorithm (SMARF) on CBCT detection of buccal fenestration and dehiscence of varying sizes adjacent to zirconia versus titanium implants. Methods: In this ex vivo pilot study, two 4 mm × 12 mm implants (one titanium and one zirconia) were implanted in a dry human mandible. Artificial fenestration and dehiscence defects were created at the buccal surface of right and left premolar regions, in three sizes (small 2 mm × 2 mm× 1 mm; medium 2 mm × 4 mm× 1 mm; large 2 mm × 6 mm× 1 mm). Scans were acquired with a Papaya 3D (Genoray) CBCT using a single-tooth FOV (4 cm× 4 cm × 5 cm), 83 kVp, 10 mA, 0.1 mm voxel; each condition was imaged with MAR on and off (six repeats each). Two trained observers scored defect presence on a five-point Likert scale. Sensitivity, specificity, inter- and intraobserver agreement (Cohen’s kappa), and comparisons across conditions were calculated. Results: Interobserver κ = 0.65 (substantial); intraobserver κ = 0.87–1.00 (excellent). For titanium implants, sensitivity and specificity were 100% across all defect types, sizes, and MAR settings. For zirconia implants, fenestration sensitivity with MAR off increased with size (small 50%, medium 75%, large 100%); MAR on raised sensitivity to 100% for all sizes, with a significant improvement for small fenestrations (p = 0.005). Dehiscence sensitivity for zirconia was 100% across sizes and MAR conditions. Specificity for zirconia defects was 83.3% with MAR off and decreased to 66.6% with MAR on (p = 0.07). Conclusions: With the Papaya 3D system used in this pilot study, MAR was unnecessary for defect detection adjacent to titanium implants but affected performance for zirconia implants: MAR increased sensitivity for small fenestrations while reducing specificity (although not statistically significant). Clinicians should weigh the sensitivity–specificity trade off when applying MAR around zirconia implants. Full article

Artificial intelligence-assisted food detection is increasingly moving from single-task classification toward integrated analytical systems capable of producing multiple quality-related outputs from one sensing workflow. However, most hyperspectral food detection studies still rely on a single spectral range or simple feature concatenation, which limits their ability to exploit complementary physicochemical information from heterogeneous sensors. In this study, an artificial intelligence-enabled visible–near-infrared and short-wave infrared (VNIR-SWIR) hyperspectral fusion framework is proposed for multi-task fruit detection, using origin authentication and bruise localization as representative tasks. The proposed method first constructs an observation-consistent fused representation from high-resolution VNIR images and low-resolution SWIR images. Collaborative spectral unmixing is used to couple cross-modal material distributions, while abundance-consistency and downsampled observation-consistency constraints are introduced to estimate SWIR-informed features on the VNIR spatial grid without assuming measured high-resolution SWIR ground truth. The fused representation is then processed by a shared spectral–spatial deep encoder with two task-specific heads: a fruit-level classification head for origin authentication and a pixel-level segmentation head for bruise detection. Experiments on apple and kiwifruit datasets show that the proposed framework outperforms VNIR-only, SWIR-only, bicubic-fusion, CNMF-style fusion, and TV-regularized fusion baselines under five fruit-level stratified random splits. For origin-category authentication, the proposed method achieved an accuracy of almost 93.85 for apples and almost 94.35 for kiwifruit. For bruise localization, the proposed method achieved higher overall accuracy, average accuracy, and Cohen’s kappa across the evaluated fruit categories. Full article

Accurate automated interpretation of upper-extremity musculoskeletal radiographs remains challenging because fracture appearance varies across anatomical regions and can be subtle under class imbalance. This study proposes a two-stage deep learning framework for MURA-based X-ray analysis, aiming to improve body-part recognition and body-part-wise abnormality detection. Multiple architectures were first compared for seven-class body-part classification, after which the selected hybrid Xception-Swin model was fine-tuned for abnormality detection within each anatomical subset. The framework combines Xception-derived local structural features with Swin Transformer contextual features using attention-based fusion, and performance was evaluated using accuracy, F1-score, AUC-ROC, Cohen’s kappa, calibration, component-level ablation, post hoc explainability, and zero-shot FracAtlas validation. For body-part classification, the model achieved accuracy = 0.9643, macro F1 = 0.9574, AUC-ROC = 0.9963, and kappa = 0.9579. For abnormality detection, accuracy ranged from 0.7289 to 0.8538, F1 from 0.7191 to 0.8508, AUC from 0.7693 to 0.9080, and kappa from 0.4449 to 0.7071. Ablation on hand and humerus radiographs showed the highest macro F1 with Hybrid Attention, while FracAtlas validation yielded AUC = 0.8247 and kappa = 0.5812. The results support complementary CNN-Transformer fusion and indicate preliminary cross-dataset generalizability. Implementation resources are available at Zenodo. Full article

This systematic narrative review synthesizes 42 distinct sources including peer-reviewed journal articles, selected conference papers, and policy documents to examine the role of knowledge management (KM) processes in Saudi higher education accreditation, with specific focus on the National Center for Academic Accreditation and Evaluation (NCAAA) standards. Drawing on literature published between 2005 and 2025, the review investigates how KM frameworks, including Nonaka and Takeuchi’s SECI model (socialization, externalization, combination, and internalization), may be associated with accreditation outcomes in Saudi universities. The reviewed literature suggests an association between systematic KM approaches and more effective accreditation processes; causal conclusions are not warranted given the observational and case study nature of the evidence base. Certainty of the overall evidence body is rated as low to moderate. The study reveals significant challenges, including information decentralization, inadequate training, resistance to change, and the absence of dedicated governance structures that appear to impede effective knowledge transfer during accreditation processes. A secondary sustainability coding pass identified associations between KM-driven accreditation practices and institutional sustainability, environmental sustainability, and alignment with SDG 4 (Quality Education) and SDG 16 (Strong Institutions); these findings are hypothesis-generating rather than confirmatory. It should be noted that all screening and data extraction were conducted by a sole reviewer; a post hoc validation exercise achieved Cohen’s kappa = 0.81 (95% CI: 0.72–0.90) for inclusion/exclusion decisions, providing retrospective assurance of acceptable consistency. This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines. Full article

Background: This study aims to delineate the present and projected suitable habitats for four dominant tick species in the Beijing–Tianjin–Hebei (BTH) region, providing a spatial basis for targeted tick-borne disease surveillance. Methods: We systematically reviewed the published literature and the Global Biodiversity Information Facility (GBIF) to compile tick occurrence records in the BTH region. A total of 167 geo-referenced occurrence records with verified coordinates were obtained for four dominant species: Haemaphysalis longicornis, Haemaphysalis concinna, Dermacentor silvarum, and Ixodes persulcatus. The MaxEnt model was applied with bioclimatic variables (WorldClim, 2.5 arc-min), elevation, slope, aspect, and NDVI. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC) for within-species comparisons, complemented by the True Skill Statistic (TSS), Cohen’s kappa, and omission rate. Future projections (2021–2040, 2041–2060, 2061–2080, 2081–2100) were made under the SSP245 scenario using only climate variables, as the NDVI and topographic variables cannot be reliably forecast. Results: The four dominant tick species showed distinct distribution patterns: Hae. longicornis was widely distributed across the BTH region, whereas Hae. concinna, D. silvarum, and I. persulcatus were mainly found in the northern and northwestern mountainous areas. The primary environmental drivers were temperature, elevation, and the NDVI. MaxEnt models achieved good predictive performance (test AUC: 0.86–0.91; TSS: 0.72–0.88). Under future climate scenarios, suitable habitat centroids were projected to shift northwestward for Hae. longicornis (~57.6 km), D. silvarum (~71.1 km), and I. persulcatus (~50.0 km), and northeastward for Hae. concinna (~63.0 km) by 2081–2100. Conclusions: In this study, we identified current and future high-risk areas for four dominant tick species in the BTH region, providing a reproducible foundation for surveillance. Future projections should be interpreted with caution as they only account for climatic changes. Full article

(1) Background: Clinical trial data extraction from registries such as ClinicalTrials.gov remains labor-intensive and error-prone, often missing critical details hidden in unstructured protocol descriptions. Large Language Models (LLMs) offer potential to automate this process, yet systematic multi-model comparisons on real clinical trial data remain scarce. (2) Methods: Four LLMs (OpenAI o4-mini-high, Anthropic Claude-Sonnet-4, Google Gemini 2.5-Pro, and Meta Llama-4-Maverick) extracted brain stimulation parameters from 67 transcranial direct current stimulation (tDCS) trials in Parkinson’s disease via a structured JSON schema. Pairwise inter-model agreement was quantified with Cohen’s Kappa and percentage agreement across binary, categorical, and multi-component task tiers. (3) Results: Under exact-string matching, agreement was near-perfect for binary classifications (non-invasive classification: 100%; brain stimulation presence: 99.3%, $\kappa$ = 0.50) and substantial for categorical extractions (primary stimulation type: 96.4%, $\kappa$ = 0.70), but fell to 48.6% ($\kappa$ = 0.43) for complex anatomical targets. Numeric parameters revealed model-specific strengths: o4-mini-high and Claude-Sonnet-4 achieved perfect duration agreement (r = 1.000, n = 19) while Llama-4-Maverick diverged substantially (r < 0.12). Validation against an expert gold standard (100% inter-annotator agreement on a 20-trial overlap) confirmed high extraction accuracy across all features (mean 93.7–98.9%). Crucially, the low agreement on anatomical targets proved to be an artifact of exact-string scoring: under the same semantic matching used to measure accuracy, inter-model agreement rose to 97.0%, coinciding with the 95.5% expert accuracy. Inter-model agreement therefore tracks accuracy once both are measured on a common basis. (4) Conclusions: Exact-string inter-model agreement decreases with task complexity, but this decline largely reflects interchangeable free-text wording rather than reduced accuracy. Evaluated semantically, agreement and expert accuracy are both high and closely aligned. A residual risk is not low accuracy but the rare error shared across all models, which agreement cannot detect, and which overall accuracy can itself mask when one class dominates. These findings inform hybrid human–AI systematic review pipelines in which targeted expert oversight focuses on shared-error and minority-class detection. Full article

This paper presents “Quantifying Airline Reputation from Multilingual Online Content: Artificial Intelligence and a Practical Application in a Lightweight Reputation-Intelligence Framework.” Airline reputation is increasingly shaped by multilingual digital narratives that evolve faster than conventional survey cycles, creating a need for timely and interpretable monitoring tools. This study develops and evaluates a lightweight reputation-intelligence framework that integrates brand-safe retrieval, multilingual transformer-based sentiment inference, zero-shot natural-language-inference aspect categorization, TF–IDF/KMeans topic induction, and short-horizon forecasting. The framework formalizes document-level outputs into managerial indicators, including a Net Sentiment Index, polarity shares, aspect scores, topic summaries, and projected sentiment trajectories. On a 3990-document annotated sentiment subset, the multilingual transformer achieved 0.9015 accuracy, 0.9048 macro-F1, 0.9050 weighted-F1, a Cohen’s kappa of 0.8492, and a Net Sentiment Index of 48.5%, while errors were concentrated between adjacent polarity classes. Aspect evaluation showed that supervised in-domain learning substantially outperformed zero-shot inference, clarifying the trade-off between cold-start portability and benchmark accuracy. The results support the framework as a pilot decision-support architecture for airline reputation sensing rather than as a definitive large-scale benchmark. The approach offers scalable and CPU-friendly monitoring capability for airlines, airports, consultants, and public-sector users, while future work should expand multilingual annotation and domain adaptation. Full article

Objectives: Sudden cardiac death (SCD) is a leading cause of mortality, accounting for approximately 50% of all cardiovascular deaths and 20% of all-natural deaths in Western countries. In individuals over 50 years of age, coronary artery disease (CAD) is responsible for more than 80% of cases, whereas in younger subjects SCD is more frequently associated with non-ischemic myocardial diseases, including hypertrophic cardiomyopathy (HCM), arrhythmogenic cardiomyopathy (ACM), dilated cardiomyopathy (DCM), and myocarditis. Additional causes in young adults include coronary artery anomalies and primary arrhythmic disorders related to channelopathies. This study evaluated the diagnostic performance of post-mortem cardiac magnetic resonance imaging (PM-CMR) in identifying morphological substrates underlying SCD in formalin-fixed explanted hearts, with particular attention to the concordance between PM-CMR findings and autopsy results in cases of sudden coronary death. Material and Methods: We retrospectively reviewed 110 PM-CMR examinations from the Regional Register of Sudden Cardiac Death of Friuli-Venezia Giulia, of which 101 were included in the final analysis. Results: PM-CMR detected pathological findings in 60 hearts (59%), including acute ischemic lesions in 39 cases and other conditions, such as hypertrophic cardiomyopathy, chronic fibrotic ischemic changes, and adipose metaplasia in 21 cases. A good agreement between PM-CMR and autopsy findings was observed (Cohen’s kappa = 0.8). Conclusions: Overall, PM-CMR proved effective in identifying relevant morphological and signal alterations, supporting conventional autopsy. Despite some limitations, particularly in hyperacute ischemic lesions, PM-CMR appears to play a promising role in the diagnostic work-up of SCD and in supporting family screening programs for primary prevention. Full article

Background/Objectives: Self-medication (SM) is a global public health phenomenon with both benefits and risks. Evidence on its determinants remains inconsistent, particularly among university students, and limited in Central and Eastern Europe despite high OTC availability and variable health literacy. This study aimed to assess SM behaviors among health sciences students and identify sociodemographic and health-related determinants of OTC use. Methods: The cross-sectional CAWI survey was conducted between November 2024 and March 2025 among 435 students of the Medical University of Warsaw selected through purposive sampling. Univariable and multivariable logistic regression analyses were performed in the study population to characterize respondents practicing SM. Detailed analyses of SM behaviors included only respondents reporting self-medication (n = 278; 64.2%). Associations between sociodemographic and health-related characteristics and selected SM behaviors were assessed using the chi-square (χ²) test, with false discovery rate correction applied using the Benjamini–Hochberg procedure. Questionnaire reliability was confirmed using Cohen’s kappa coefficients ranging from 0.63 to 1.00. Statistical significance was set at p < 0.05. Results: The most commonly used medicines were analgesics (97.5%), vitamins (54.7%), and antipyretics (48.2%). Adjusted multivariable analysis showed that the odds of SM were higher among women (AOR = 4.11, 95% CI = 2.11–8.02), Emergency Medical Services students (AOR = 3.57, 95% CI = 1.45–8.75), master’s degree students (AOR = 3.45, 95% CI = 1.96–5.88), and students living in large cities with more than 500,000 inhabitants (AOR = 2.63, 95% CI = 1.43–5.00). Motivations, perceived benefits, risks, and adherence to package leaflet instructions differed significantly by respondent characteristics. Conclusions: SM behaviors are influenced by sociodemographic and health-related factors. Targeted education on rational OTC use, professional responsibility, and critical appraisal of health information is needed. Safe SM practices among health sciences students may support their future role in promoting responsible self-medication. Full article

Background: Abdominal pain in hematological patients, particularly during chemotherapy-induced neutropenia, represents a significant diagnostic challenge due to the broad spectrum of potentially life-threatening conditions, including neutropenic enterocolitis (NEC). Computed tomography (CT) is considered the reference imaging modality; however, its use is limited by radiation exposure, and the need for patient transport. Bedside ultrasound (BS-US) may offer a rapid, non-invasive, and repeatable alternative. Methods: This prospective study compared BS-US and CT in 65 hematological patients presenting with acute abdominal pain. Concordance between the two modalities was evaluated in terms of intestinal site localization, bowel wall thickness (BWT), and final diagnosis. Diagnostic agreement was assessed using Cohen’s kappa coefficient, and additional diagnostic accuracy metrics—including sensitivity, specificity, positive predictive value, and negative predictive value—were calculated. BWT measurements were analyzed using Bland–Altman methods. Results: A high level of agreement was observed between BS-US and CT in both intestinal localization and final diagnosis. Agreement for intestinal site localization was good (Cohen’s κ = 0.964), as was diagnostic concordance (Cohen’s κ = 0.962), and using CT as the reference standard, BS-US showed uniformly good diagnostic performance across all evaluated conditions, with sensitivity, specificity, PPV, and NPV consistently reaching 1.00 and confirming strong agreement between BS-US and CT. These findings were consistent across different clinical settings (hematology unit and Intensive Care Unit) and independent of body mass index. In NEC cases, BWT measurements showed strong concordance between CT and BS-US, with only 4.6% of values outside the limits of agreement in Bland–Altman analysis. Conclusions: BS-US demonstrated a good agreement with CT and proved to be a reliable, safe diagnostic tool in hematological patients with acute abdominal pain. These findings indicate that bedside ultrasound represents a valuable and safe diagnostic tool in neutropenic hematological patients with acute abdominal pain, providing crucial information in a clinically fragile population that may not always be suitable for CT due to their unstable condition. While our study is hypothesis-generating, the role of BS-US in this setting emerges as a reasonable, evidence-supported hypothesis that warrants further prospective evaluation. Full article

Selective Sustainable Development Goal (SDG) disclosure is frequently interpreted as evidence of weak university accountability. This research advances a different position. It argues that the problem is not selectivity itself, but unexplained selectivity. In plural-mission institutions, no university can give equal depth to every SDG without risking superficiality. Selectivity preserves accountability when priorities are publicly justified, mission-linked, evidence-supported, and explicit about what remains underdeveloped. It erodes accountability when curated visibility replaces public account-giving. Drawing on structured document analysis of 107 university SDG and sustainability disclosures from Europe, Asia–Pacific, North America, Latin America, Africa and the Middle East, published between 2019 and 2024, independently coded by two researchers (Cohen’s kappa = 0.81) and scored through the SDG Disclosure Review-Readiness Index (SDG-DRRI), the research shows that moderately selective documents (n = 18) achieved the highest mean score, outperforming both broad reporting (n = 34) and highly selective reporting (n = 55). This research contributes a new distinction between accountability-preserving and accountability-eroding selectivity, showing why disclosure breadth alone is an inadequate proxy for accountability quality. It advances sustainability accounting and higher education reporting debates by reframing SDG selectivity as a conditional accountability phenomenon: legitimate when publicly justified, problematic when it narrows responsibility without explanation. Full article

Layered cloud cover (LCC) describes the vertical distribution of cloud occurrence and is a key variable for assessing the radiation budget of the Earth-atmosphere system. However, ground-based radars have limited spatial coverage, while existing satellite cloud-cover products rarely provide both spatiotemporal continuity and high accuracy. Because nighttime satellite observations lack visible-channel information, conventional passive satellite remote sensing remains limited in providing day-night continuous LCC retrievals. In this study, we propose an infrared-based framework for retrieving large-scale day-night LCC from geostationary satellite observations. The framework first resolves cloud vertical structure using a hybrid machine learning and physical algorithm for day-night cloud-base height (CBH) retrieval, and then derives cloud cover in different vertical layers. Validation against active measurements from spaceborne and ground-based cloud radar demonstrates that the satellite-retrieved LCC captures cloud vertical distributions and their diurnal variations. The cloud-layer identification accuracies reach 76.3% and 77.9% for daytime and nighttime, respectively, with corresponding Cohen’s kappa coefficients of 0.66 and 0.68. The primary source of algorithmic uncertainty is the low precision of low-cloud identification, which is constrained by objective factors and physical characteristics. The retrieved annual mean LCC fields reproduce major climatological features, including enhanced high and deep convective clouds over the tropical western Pacific and dominant low-cloud occurrence over the mid-latitude oceans. A case study of Typhoon Doksuri further shows that the 10 min LCC retrievals capture the vertical evolution of the typhoon cloud system during intensification, eyewall structural adjustment, landfall, and post-landfall decay. These results indicate that the proposed infrared-based retrieval framework provides a promising basis for constructing large-scale day-night LCC datasets and can support cloud-radiation studies, climate-model evaluation, and weather monitoring. Full article