Atmospheric and Oceanic Sciences

[This corrects the article DOI: 10.1371/journal.pone.0335141.].

While data assets are increasingly recognized as strategic resources in the digital economy, their implications for financial reporting quality remain underexplored. Employing a textual analysis approach based on deep learning techniques to construct a measure of data assets, this study investigates their impact on real earnings management (REM) using a sample of Chinese A-share listed firms from 2010 to 2023. We document a negative association between data assets and REM. Our findings remain robust after addressing endogeneity concerns using instrumental variables based on regional digital infrastructure and propensity score matching. Mechanism analyses reveal that data assets mitigate REM through two distinct channels: curbing managerial short-termism and strengthening external monitoring via increased analyst coverage. Cross-sectional tests indicate that this governance effect is more pronounced among firms facing earnings pressure. Furthermore, data assets primarily constrain cost-side manipulation strategies, specifically abnormal production costs and discretionary expenditures. Finally, we show that the reduction in REM driven by data assets contributes to enhanced long-term firm value. Collectively, these findings uncover a novel governance effect of data assets, suggesting that digital infrastructure investments yield positive externalities for corporate financial behavior.

Medical intelligent question-answering (QA) systems have become important tools for improving the efficiency of healthcare services, and recent research has increasingly emphasized performance optimization and multimodal integration. However, existing systems still face several challenges in intent recognition, entity extraction, and multimodal knowledge fusion, particularly reduced accuracy in multi-label classification, heavy reliance on large-scale annotated data, and limited support for cross-modal retrieval. To address these issues, this study proposes a medical intelligent QA framework that integrates a dual-layer attention mechanism, a large language model, and a multimodal medical knowledge graph to improve system understanding and response generation in complex clinical scenarios. Specifically, we develop a text-based intent recognition model with a dual-layer attention architecture, in which a global contextual attention module is introduced to capture long-range semantic dependencies and improve multi-label classification performance. In addition, an instruction-tuned large language model is employed for zero-shot medical entity recognition, thereby reducing dependence on manually annotated datasets. Building on this foundation, we construct a multimodal medical knowledge graph comprising more than 15,000 associated medical images and develop a visualization-oriented retrieval interface using Flask and ECharts. Experimental results show that the proposed intent recognition model achieves a peak Micro-F1 of 94.42% on multiple benchmark datasets, outperforming several baseline methods. The LLM-based entity recognition module achieved competitive recall in medical entity extraction, demonstrating strong capability in identifying medical entities. User evaluation results further indicate that the system is effective and practical across a variety of medical query types. This study provides a feasible framework for advancing medical QA systems through improved intent recognition, low-resource entity extraction, and multimodal knowledge integration.

This study examined the acute effects of high-load resistance training and low-load resistance training with blood flow restriction on sprint ability, pennation angle and lower limb muscle stiffness. A randomized crossover trial was conducted in 17 collegiate male sprinters. Four intervention conditions were applied: high-load (90% 1RM hip thrust [HT] or half-squat [HS]) and low-load (30% 1RM HT or HS with blood flow restriction [BFR]) protocols. Muscle-tendon stiffness, pennation angle, and 20-m sprint performance were assessed at 5 minutes post-intervention. The results showed that statistically significant differences were observed among interventions (p < 0.05). Regarding sprint performance, the 30% HT + BFR significantly improved 0-20m sprint time, whereas the 30% HS + BFR significantly improved 10-20m sprint time compared to high-load conditions. In terms of muscle architecture, both 30% HS + BFR and 90% HS conditions elicited a significant decrease in the pennation angle of the rectus femoris. Moreover, muscle-tendon stiffness generally increased following conditioning activities; notably, the 30% HT + BFR condition increased rectus femoris stiffness, while the 30% HS + BFR condition significantly increased stiffness in the gastrocnemius and Achilles tendon compared to other interventions or baseline. In conclusion, this study demonstrates that low-load resistance exercise combined with blood flow restriction produces greater acute improvements in sprint performance in male sprinters than high-load training, accompanied by reductions in pennation angle and increases in muscle-tendon stiffness.

INTRODUCTION: Academic stress (AS) can be understood as a multidimensional form of chronic stress that has been linked to adverse patterns across physical and mental health in university students. This protocol outlines a longitudinal, observational study to examine within-semester associations between AS and indicators of nutritional status, eating-related patterns, sleep quality, and depressive symptoms. By integrating physiological, behavioral, and psychological indicators within a repeated-measures framework, the study aims to generate evidence that can inform prevention-focused actions and early support strategies in university health contexts. HYPOTHESES: (i) Higher AS will be longitudinally associated with less favorable eating-related patterns, nutritional and somatic indicators, and higher depressive symptoms in university students. (ii) Higher AS will be associated with altered post-awakening cortisol dynamics, specifically a lower cortisol awakening response (CAR) and lower AUCi, together with a higher AUCg, and with higher salivary alpha-amylase (sAA) levels, consistent with stress-related activation of HPA-axis and sympathetic pathways. DESIGN AND PARTICIPANTS: A longitudinal, observational, nonexperimental design with repeated measurements will be implemented across three assessment cycles during an academic semester. A feasibility-based convenience sample will be recruited from undergraduate students (2nd to 4th year) enrolled in the Faculty of Medicine, University of Concepción (Chile). Students receiving psychological or pharmacological treatment will be eligible to reflect real-world heterogeneity and support ecological validity. METHODS: Data will be collected through standardized questionnaires, nutritional assessments, biological sampling, and wearable-derived somatic indicators. Electronic surveys administered via REDCap will assess AS, perceived stress, eating-related patterns, and depressive symptoms. Diet will be assessed through interviewer-administered nutritional interviews, including repeated 24-hour dietary recalls treated as time-specific observations and modeled longitudinally as time-varying measures, and complemented by diet-quality and dietary inflammatory potential indices. Wearable devices will record nonclinical somatic indicators, including heart rate, oxygen saturation, and sleep-related metrics during monitoring periods. Saliva will be collected twice per week during each assessment cycle to quantify salivary cortisol dynamics and sAA activity, and peripheral blood samples obtained at baseline and end of semester will be used to determine lipid profile, fasting glucose, albumin, globulin, and total proteins. STATISTICAL ANALYSIS: Analyses will include descriptive and bivariate summaries, followed by multivariable models appropriate to outcome type. Longitudinal associations will be examined using mixed-effects models, and temporal cross-lagged associations will be explored using random-intercept cross-lagged panel models across the three assessment cycles. All inferences will be framed as associational given the observational design. EXPECTED RESULTS: Rather than prespecifying outcomes, this protocol is expected to generate longitudinal evidence on how within-semester variation in AS aligns with eating-related patterns, diet, nutritional and somatic indicators, depressive symptoms, and stress-related physiological indicators (salivary cortisol dynamics and sAA activity). The integrated, feasibility-oriented measurement framework may serve as a replicable template for future research and inform prevention-focused actions to support student well-being in university settings.

Kenya's socioeconomic development is heavily dependent on the construction industry; however, poor risk and contract management typically results in cost overruns, delays, and quality deficiencies that undermine project performance. Therefore, this study examined the way risk management practices affect the performance of building projects in the Nairobi Metropolitan Area, considering contract management practices and the moderating effect of project managers' skills. The Theory of Constraints and Agency Theory, which place a strong emphasis on supervision, responsibility, and the removal of performance barriers, serve as the foundation for the analysis. The study used a descriptive and explanatory research design. 127 completed construction projects (95 residential and 32 non-residential) comprised the target population for this study. From this, 64 valid responses (96% response rate) were obtained from a sample of 67 projects chosen by stratified random sampling using Slovin's technique at a 95% confidence level. A standardized Likert-scale questionnaire was used to collect data. Inferential statistics, such as simple linear and hierarchical multiple regressions and Pearson's correlation analysis, were utilized to evaluate relationships and prediction effects. Descriptive statistics, such as means, percentages, and standard deviations, were utilized to define the project's characteristics. Expert review was used to guarantee the validity of the instrument, and reliability was verified using Cronbach's alpha coefficients greater than 0.70. The results demonstrate that both risk management practices (β = 0.612, p < 0.001) and contract management practices (β = 0.239, p = 0.020) are significant positive predictors of construction project performance, with the combined model accounting for 60.6% of the variance in construction project outcomes (R² = 0.606). Project manager skills did not significantly moderate the relationship between these management practices and project performance (interaction terms p > 0.05) and exhibited no significant independent effect (β = 0.132, p = 0.120), contributing only a marginal increase in explanatory power to R² = 0.62. The study concludes that while robust risk and contract management frameworks are critical for project success, the additive value of project manager skills remains modest and non-moderative. Recommendations include establishing uniform industry protocols and continued emphasis on professional certification and leadership development to support, rather than substitute systemic management practices.

BACKGROUND: The hemoglobin-to-red cell distribution width ratio (HRR) is a biomarker associated with systemic inflammation and outcomes in critical illness. Within clinical databases, there exists an extreme-prognosis subgroup of inflammatory bowel disease (IBD) patients, those who all died within 30 days of ICU admission. Due to limitations in the database, this study can only analyze this specific subgroup. OBJECTIVE: This study aims to explore and describe the association between admission HRR and the occurrence of sepsis during ICU stay in this specific subgroup of IBD patients. METHODS: A retrospective cohort study was conducted using the eICU Collaborative Research Database (2014-2015), including 229 eligible patients. Multivariable logistic regression was used to assess the independent association, adjusting for confounders. The dose-response relationship was examined using restricted cubic spline (RCS) models. Subgroup and interaction analyses were performed across age, sex, and race. RESULTS: The sepsis group had a significantly lower admission HRR than the non-sepsis group (6.04 ± 1.66 vs. 6.77 ± 2.0, P = 0.008). After full adjustment, each 1-unit increase in HRR was associated with 20.3% lower odds of sepsis (P = 0.007). Compared to the lowest quartile (HRR < 5.1), patients in the highest quartile (HRR ≥ 7.85) had 66.8% lower adjusted odds of sepsis (P = 0.015). RCS analysis indicated a linear, inverse relationship between HRR and sepsis (P for non-linearity = 0.42). Subgroup analysis revealed the negative association was more pronounced in patients aged <65, males, and White patients. However, formal interaction tests were not statistically significant (all P for interaction >0.05), indicating the association did not differ meaningfully across these subgroups. CONCLUSIONS: In this specific cohort, a lower admission HRR was associated with the occurrence of sepsis. Due to inherent selection bias, this finding describes an association within this specific subgroup and cannot be generalized. This exploratory study generates the hypothesis that HRR may reflect a unique pathophysiological state in end-stage IBD, a hypothesis that requires validation in prospective, unbiased cohorts.

AIMS: To evaluate plasma neurofilament light chain (NfL) and glial fibrillary acidic protein (GFAP) as candidate biomarkers in myasthenia gravis (MG). METHODS: Ninety MG patients and 40 healthy controls were recruited. Disease severity was assessed by the Myasthenia Gravis Foundation of America (MGFA) classification, Myasthenia Gravis Composite (MGC) score, and Myasthenia Gravis Activities of Daily Living (MG-ADL) scale. Plasma NfL and GFAP were quantified using Single Molecule Array (Simoa) assays. RESULTS: NfL and GFAP plasma concentration did not differ between MG and controls (p > 0.05). Neither biomarker correlated with MG-ADL or MGC, and no differences were observed across MGFA classes (p > 0.05). Biomarker levels were unrelated to myasthenic crisis history or treatment exposure. CONCLUSION: Plasma NfL and GFAP, although informative in other neuroimmunological and neurodegenerative conditions, do not distinguish MG from healthy controls and show no association with disease severity. This study adds to the emerging literature on NfL in MG and represents one of the larger controlled analyses incorporating both NfL and GFAP biomarkers in this disease. The findings argue against adopting NfL or GFAP for MG monitoring and highlight the need for MG-specific biomarker strategies.

Inflammation is an inherent feature of advanced chronic kidney disease (CKD) and associated with adverse outcomes. Several inflammatory biomarkers have been shown to be associated with mortality in CKD. Cell-free DNA (cfDNA) is a novel biomarker for inflammation which has not been previously examined in patients with CKD stage 4-5 not undergoing dialysis. cfDNA was extracted from plasma and quantified with Qubit Flex Fluorometer using dsDNA High Sensitivity kit in 138 patients with CKD stage 4-5 not undergoing dialysis at baseline and at a control time point of median 2.7 years of follow-up. A ratio of control and baseline measurement adjusted for an increment of one year of follow-up was calculated. Associations between cfDNA at baseline and all-cause mortality, major adverse cardiovascular and cerebrovascular events (MACCE, defined as a composite outcome of acute myocardial infarction, coronary revascularization, ischemic or hemorrhagic stroke and cardiovascular death), emergency room (ER) visits, hospitalizations or incident malignancies were assessed. Within a median follow-up of 6.2 years, no associations were observed between cfDNA and mortality, MACCEs, ER visits, hospitalizations or incident malignancies. cfDNA control measurement and cfDNA delta ratio was available in 101 patients. Patients who had received a kidney transplant by the control cfDNA measurement had significantly higher cfDNA delta ratio compared to patients not on renal replacement therapy (RRT) and those undergoing dialysis (p < 0.001 for both comparisons). Patients undergoing dialysis at the time of the control cfDNA measurement had higher cfDNA delta ratio (p = 0.003) compared to patients not on RRT. The present study is the first to show that cfDNA is not associated with adverse outcomes in patients with advanced CKD not undergoing dialysis at baseline. Furthermore, our results provide unique data on the evolution of cfDNA levels in predialysis CKD stage 4-5 patients transitioning to different modalities of RRT.

AIM: This study investigated the impact of a structured undergraduate nursing research course on Saudi nursing students' attitudes toward nursing research. METHODOLOGY: A quasi-experimental pretest-posttest design was used to assess the influence of a 14-week nursing research course at a public Saudi university. Attitudes were measured with a validated questionnaire administered to 62 students before and after the course. A paired t-test was used to evaluate the difference in students' attitudes toward nursing research. RESULTS: Students' attitudes toward nursing research increased significantly after the course (M = 69.58, SD = 9.91) compared to pre-course levels (M = 66.28, SD = 8.43); t(55) = -2.49, p = 0.01). Notable gains occurred in attitudes toward research skills (p = 0.01) and the use of research in clinical practice (p = 0.007). However, there were no statistically significant differences in personal interest in research or in its perceived usefulness. Male students with lower GPAs demonstrated the largest positive attitude shifts. CONCLUSION: The research course chiefly enhanced attitudes toward research skills and clinical application, without raising interest or perceived usefulness of nursing research among undergraduate nursing students. Educators should use diverse teaching methods to promote students' interest in and appreciation of nursing research.

INTRODUCTION: Although various bracket recycling methods have been studied, the combined influence of bracket reconditioning methods and primer protocols on the bonding performance of debonded orthodontic brackets remains unclear. This study aimed to evaluate the effects of three reconditioning methods and three primer protocols on the bond strength of rebonded orthodontic brackets to enamel. MATERIALS AND METHODS: One hundred stainless steel brackets were bonded to extracted premolars and divided into 10 groups (n = 10). Ninety brackets were debonded, reconditioned using OneGloss bur, flaming, or sandblasting, and rebonded with Mani Bond, Denu Bond, or no primer. Ten new brackets served as controls. Bracket bases were examined using scanning electron microscopy (SEM). Bond strength, adhesive remnant index (ARI), and bonding reliability were evaluated using SBS testing, ARI scoring, and Weibull analysis. RESULTS: Bond strength was significantly influenced by the interaction between reconditioning method and primer protocol (p < 0.001). Sandblasting combined with Mani Bond produced the highest bond strength (9.33 ± 3.02 MPa), whereas OneGloss and sandblasting combined with Denu Bond achieved bond strengths comparable to new brackets and higher reliability, reflected by greater Weibull moduli (4.74 and 4.72). SEM demonstrated more effective mesh cleaning after sandblasting than after OneGloss bur treatment or flaming. ARI distributions differed significantly among groups (p = 0.01), with failures occurring predominantly at the bracket-adhesive interface. CONCLUSIONS: The bonding performance of rebonded orthodontic brackets is influenced by the interaction between reconditioning and priming protocols. Appropriate combinations of these procedures may improve the rebonding outcome.

Adoption of improved crop technologies is widely recognized as essential for enhancing agricultural productivity, yet their validation under real farming conditions remains limited. The scaling trial was aimed to provide an interesting agronomic challenge through introducing a new technology that contradicts the established local field production practice. The trial was conducted over four consecutive production years (2019-2022) involving 109 (22 female) purposively selected farmers who managed a collective 32.5 ha of farmland. The farmers were strategically selected to promote the dissemination of new technologies, create demand, strengthen stakeholder linkages, and establish a sustainable technology multiplication system in the area. The comprehensive quantitative and qualitative data were collected and analyzed using proper statistical methods. The results demonstrated that the improved field pea technology was provided a 71.1% yield advantage compared to the existing local practices. With continuous expert support, 75% of farmers applied the full technology package, although 62.5% perceived it as labor-intensive, particularly during planting and thinning stages. Despite this, 95.9% of participants expressed a strong interest in cultivating the new technology in the future. Farmer-to-farmer diffusion was evident, with 1520 kg of improved seed distributed by 75% of involved farmers and stakeholders to non-participant but interested farmers. The farmers and stakeholders who attended different field days were also impressed and committed to adopting the new technology, recognizing its potential adaptation to moisture-deficient areas. These findings hence confirm the agronomic and social viability of the new technology. Therefore, it is recommended that scaling up and out of this improved field pea technology in similar areas be facilitated by the establishment of viable seed-multiplying cooperatives and strengthened stakeholder linkages.

Editorial Note: Field and laboratory assessment of larvicidal activity of tobacco plants and the cigarette butt waste on Culex pipiens (Linnaeus, 1758), Aedes aegypti (Linnaeus, 1762) L. and non-target organisms

[This corrects the article DOI: 10.1371/journal.pone.0348681.].

The Segment Anything Model 2 (SAM 2) has emerged as a robust foundational model for video object segmentation. Nevertheless, it exhibits notable limitations in crowded scenarios, particularly those involving fast-moving objects or self-occlusion. Furthermore, its greedy-selection memory architecture is afflicted by "error accumulation," which collectively impairs its performance in consistent object segmentation tasks. To address these critical drawbacks and enhance the consistency and robustness of SAM 2 in complex video scenarios, an enhanced variant of SAM 2, termed UAMP, is proposed, which integrates memory propagation based on explicit appearance and motion uncertainty modeling, coupled with a dual mechanism encompassing long-term memory updating and short-term memory selection. By fusing uncertainty-aware representations with these dual memory mechanisms, UAMP effectively accommodates dynamic variations in object appearance and motion, further refines the object memory bank, and thereby realizes consistent video object segmentation. Quantitative and qualitative evaluations conducted on diverse benchmark datasets demonstrate that UAMP achieves superior performance, particularly in scenarios involving occlusions and object reappearances. Specifically, UAMP yields consistent performance improvements over state-of-the-art (SOTA) methods across five video object segmentation (VOS) benchmarks, with a maximum enhancement of 5.6 points in the J&F metric. These findings underscore the robustness and effectiveness of the proposed UAMP method in addressing complex tracking scenarios, thus providing a valuable enhancement to SAM 2 for practical video object segmentation applications.

Mangrove forests along the Iranian coast of the Persian Gulf are valuable both economically and for biodiversity. The main tree species in these forests is Avicennia marina (Forssk.) Vierh. The habitats of this species in Iran are separated by several thousand kilometers from other populations worldwide. Although introgressive hybridization is common in this species, no morphological study has evaluated this issue in Iranian populations. This study investigated the morphological variation of Iranian A. marina and used sequencing of ITS regions to infer phylogenetic relationships Iranian A. marina populations and other Avicennia species. Morphological analyses based on pollen grains and internodes have shown differences between A. marina populations which is the main species of the mangrove forests of Iran and a new population was identified, which is here described as A. marina subsp. australis. The phylogenetic analysis based on the ITS region confirmed the monophyly of Iranian populations and secondary analysis of ITS2 has moderate support from the two populations detected in the present study. Divergence time estimates suggest these lineages separated during the Pliocene-Pleistocene transition, likely due to climatic fluctuations and sea-level changes.

Abstract. Studying the pathways of atmospheric moisture and heat is crucial for understanding global water and energy cycles, and their response to climate change. Here, we present a new global dataset of atmospheric parcel trajectories generated with the FLEXible PARTicle dispersion model (FLEXPART v11) and forced by ERA5 reanalysis. The dataset spans 1979–2024 and provides a consistent and physically grounded record for studying Lagrangian moisture and heat transport. The dataset includes 20 million global, domain-filling air-parcel trajectories together with their (thermo)dynamic properties, enabling detailed investigation of long-range atmospheric transport processes. By providing the complete trajectory archive openly, the dataset enables quantitative analyses of moisture and heat pathways without the need to perform computationally expensive Lagrangian simulations. While the trajectory dataset itself can be used with any moisture and heat tracking attribution methodology, here it is explored using the new version of the Heat And MoiSture Tracking framEwoRk (HAMSTER v2). The dataset's usability is demonstrated by (i) global analyses of moisture source–sink patterns and recycling over multiple decades, (ii) global attribution of diabatic temperature increments to upwind surface sensible heat fluxes for a representative year (2021), and (iii) two local-scale case studies which showcase how the dataset and associated tools can be applied to hydrological and temperature extremes across a range of spatial and temporal scales. Overall, this resource lowers computational barriers and supports reproducible research across the atmospheric science community. The dataset is available at https://doi.org/10.5281/zenodo.17952362 (Deman et al., 2025).

Fine particulate matter (PM 2.5 ) from industrial sources remains a major health concern. However, current emission controls and source apportionment models ignore condensable particulate matter (CPM). Here, we provide the first nationwide quantification of industrial CPM contributions to atmospheric PM 2.5 in China through an integrated framework of field measurements, receptor modeling, and chemical transport simulations. Industrial CPM exhibits a chemical profile dominated by sulfate and ammonium, distinct from filterable PM, which creates a risk that its contribution may be misattributed to secondary aerosols in traditional source apportionment. Across industrial sources, CPM mass proportions vary significantly ( p < 0.01) for NH 4 +, NO 3 –, and Cl – . Based on receptor modeling, CPM contributes a median of 3.6% (95% CI: 0.9%–14.5%) to urban PM 2.5 across all four seasons in 2023, in the industrialized cities examined (Shanghai, Xi’an, Hefei, Shijiazhuang). The contribution peaks in winter (9.5%, 95% Monte Carlo-derived intervals: 2.4%–23.1%) due to enhanced gas-to-particle condensation at low temperature in these cities. From 2014 to 2023, the iron and steel sector emerged as the dominant CPM source, with its contribution increasing by up to 31.0% in key regions during winter. This overlooked source appears to substantially compress the apportioned contribution of secondary inorganics (by an estimated 5.6%–16.6%), suggesting that omission of CPM may lead to overestimation of secondary inorganic aerosol in traditional source apportionment. The results underscore the priority control of CPM emissions from industrial sources.

Abstract. This paper presents the development, validation, and preliminary application of a sub-national scale crop yield emulator to be integrated into the compact Earth system model OSCAR. The emulator simulates yields for four major food crops: maize, rice (two growing seasons), soybean, and wheat (spring and winter varieties), in alignment with the Agricultural Model Intercomparison and Improvement Project (AgMIP) and the Inter-Sectoral Impact Model Intercomparison Project (ISIMIP) framework. Key drivers include atmospheric CO2 concentration (represented as C), growing season temperature (T), water availability (W), and nitrogen fertilization (N). The emulator is trained on an ensemble of process-based crop model simulations from AgMIP’s Global Gridded Crop Model Intercomparison Projects (GGCMI), which is based on the ISIMIP Phase 3 protocol. The crop models used bias-corrected historical and future climate scenarios under fixed socioeconomic conditions, to estimate yield responses under various scenarios until the end of this century. Evaluation of the emulator against the crop model outputs demonstrates the emulator's ability to replicate complex model behavior with high fidelity. Additionally, the emulator-derived yield sensitivities to CO2 and temperature are consistent with those observed in field experiments, reinforcing its empirical robustness. Historical simulations incorporating time-varying nitrogen inputs show significantly improved agreement with FAO yield statistics, underscoring the emulator’s reliability over the historical period and its potential for future impact assessments. This study provides a computationally efficient yet empirically grounded tool for representing crop yield responses, bridging the gap between complex crop models and statistic models. The developed crop emulator facilitates probabilistic projections across large ensembles of climatic and socio-economic scenarios at policy-relevant, sub-national scales. Potential applications include integrated assessments of future food security under climate and land-use change, as well as evaluations of bioenergy with carbon capture and storage (BECCS) potential from crop residues.

The rapid intensification (RI) of tropical cyclones (TCs) remains most persistent challenges in operational forecasting, particularly in western North Pacific (WNP), where RI events are most frequent and intense globally. Although numerical weather prediction continues to advance, adaptive tools are needed to resolve the multi-scale processes driving sudden intensity changes. Existing studies often rely on static binary thresholds for RI occurrence. However, the physical mechanisms favoring RI emerge at varying stages during the intensification process across different cases, governed by concurrent environmental conditions and internal vortex dynamics rather than any fixed RI definition. To address these limitations, this study establishes a diagnostic system for TC intensification in the WNP by integrating vortex-scale reanalysis data for capturing structural drivers of RI, a new continuous intensification rate (IR) index that moves beyond traditional binary RI classifications, and machine learning techniques. Following systematic hyperparameter optimization, the Random Forest, Support Vector Regression, and Artificial Neural Network models all demonstrated consistent and reliable performance in linking higher IR index values to increased probabilities of TC intensification. By integrating physically meaningful vortex-scale features with a continuous IR metric, this framework offers a versatile approach to advance the understanding and forecasting of TC intensification in the WNP.

Abstract. Real-time and accurate three-dimensional ocean temperature–salinity (T–S) field are of great significance for a deeper understanding of ocean dynamics and prediction skill improvement of numerical models. However, current ocean observations, especially those below the sea surface, still suffer from significant limitations in temporal and spatial resolution. Several neural network methods using multi-source satellite data for underwater temperature and salinity reconstruction have been proposed, achieving real-time temperature and salinity reconstruction, but their biases relative to in-situ observations are still significant. This study focuses on the northwestern Pacific region (0–40° N, 120–160° E) and proposes an attention-enhanced three dimensional U-Net++ model, which reconstructs daily T–S fields (26 layers, 1/4° resolution, 5–2000 m depth) using real-time available sea surface temperature (SST) and sea surface height (SSH) data. The model introduces cross-scale feature aggregation and selective information gating, allowing it to emphasize temporally coherent surface features most relevant to subsurface variability, while suppressing noise propagation and over-smoothing. By integrating 26 consecutive days of SST and SSH as inputs, the model effectively alleviates the underdetermined problem of mapping limited surface observations to full-depth structures. In addition, a two-stage transfer learning strategy is employed: the model is first pretrained using monthly SST/SSH data and the gridded Argo data to learn observation-dominated low-frequency spatiotemporal patterns, and then fine-tuned using daily SST/SSH data and the high-resolution reanalysis to capture the meso-scale dynamic processes. Evaluation results show that the reconstructed T–S fields agree better with in-situ T–S profiles from World Ocean Database than previous studies, both during the validation period and in long-term statistical analyses, suggesting that the proposed approach is reliable and accurate for subsurface ocean field reconstruction. The reconstructed T–S field is available at https://doi.org/10.57760/sciencedb.31950 (Wang et al., 2025).

BACKGROUND: Ambient PM2.5 exposure is strongly associated with adverse health effects, including all-cause mortality. However, the lack of monitoring networks globally necessitates a better understanding of the spatiotemporal distribution of near-surface PM2.5 pollution. While ground-level pollutants are traditionally measured at fixed stations, integrating land use and atmospheric reanalysis data captures the broad geospatial trends and specific aerosol compositions necessary for high-resolution exposure assessment. This study aims to demonstrate the accuracy and reliability of ensemble modeling for estimating near-surface PM concentrations at a high spatiotemporal resolution in Athens, Greece. METHODS: Daily PM2.5 concentrations were estimated using a stacked ensemble machine learning model to incorporate distributed random forest, gradient boosting, and feedforward neural network algorithms to minimize predictive error compared to individual models. The input set for all base learners consisted of daily observations from air quality monitors between 2007-2019 combined with satellite-derived estimates, providing a total of 61 variables describing regional aerosol, weather, and land use characteristics. RESULTS: We observed strong predictive performance in our ensemble model, with a mean R2 of 0.85 and an average error of 4.18 μg/m3. The annual average concentration of PM2.5 (22.30 μg/m3) exceeded current WHO and EU guidelines, with considerable spatiotemporal variation across greater Athens. The highest annual mean PM2.5 concentrations were in 2007 (33.18 μg/m3) and on average year-to-year, PM2.5 concentrations were highest during the mid-winter months, in agreement with the expected seasonal maximum for the region, likely driven by increased residential heating alongside winter meteorological conditions, such as temperature inversions and a shallow, stable planetary boundary layer. SIGNIFICANCE: This is among the first studies to estimate PM2.5 exposures in the greater Athens region at a high spatiotemporal resolution using diverse satellite and land use data. This framework enables the investigation of cumulative exposures, particularly in regions with limited ground-level monitoring.

Abstract Drought events are becoming more frequent and severe under global climate change, threatening food security and ecosystem stability. As a key indicator of atmospheric water cycling and precipitation formation, precipitable water vapor (PWV) is closely linked to drought occurrence and evolution. However, existing PWV-based drought monitoring approaches are mostly limited to regional applications, show weak performance at short time scales, and have rarely been extended to future forecasting. To address these limitations, a novel multi-scale drought monitoring and forecasting framework based on PWV and precipitation is proposed. First, a multi-scale precipitation conversion index (PCI) is calculated, and an XGBoost-based standardized PCI, termed SPCI-XG, is developed. Subsequently, a convolutional long short-term memory (ConvLSTM)-driven spatiotemporal forecasting model is developed for drought forecasting. Results indicate that global SPCI-XG agrees well with the standardized precipitation evapotranspiration index (SPEI) across time scales from 1 to 24 months, with a correlation coefficient (R) of 0.95 at the 12-month scale. It also captures the spatiotemporal evolution of drought like SPEI in semi-arid, tropical, and desert regions. Moreover, the spatiotemporal forecasting model achieves a root mean square error (RMSE) and an R value of 0.29 and 0.95 for single-step forecast (i.e., one-month-ahead prediction), and maintains relatively reliable predictive skill within the 1–5 month forecasting range for multi-step forecasts (i.e., forecasts for future 1–10 months) across six continents. These findings highlight the potential of PWV-precipitation coupling for globally scalable drought monitoring and forecasting, providing a new pathway for early drought warnings and climate risk assessments.

This study presents the first systematic assessment of CO 2 storage potential in the Cretaceous Abu Roash-A sandstone reservoir of the Beni Suef Field in Egypt’s Western Desert. An integrated, multidisciplinary workflow was employed, combining two-dimensional (2D) seismic interpretation, petrophysical evaluation, and 3D static reservoir modelling to characterise the structural framework, assess reservoir quality, and quantify the anticipated CO 2 storage volume. Structural interpretation reveals a fault-bounded anticlinal trap with three-way dip closures against NE-SW and NW-SE trending normal faults, suggesting geological conditions favourable for lateral containment. The reservoir shows a well-defined oil-water contact at 4,200 ft and a hydrocarbon column of 300 ft. Stratigraphic interpretation indicates a thick, regionally extensive caprock system (Khoman and Apollonia formations) overlying the sandstone reservoir. Petrophysical analysis of the Abu Roash-A sandstone indicates low to moderate shale content (Vsh up to 0.25 in northeastern zones, lower in southwestern areas), favourable porosity (15%–24%), and moderate permeability (mean ∼12.5 mD, reaching up to 150 mD) across the structural crest, identifying the elevated block between wells BS-5X and BS-11 as the optimal injection interval. Probabilistic storage calculations yield an estimated CO 2 capacity ranging from 20 Mt (low scenario) to 140 Mt (high scenario), with a base case of approximately 60 Mt, and the central and southwestern sectors showing the highest potential. The results provide a transferable framework for CO 2 storage projects in mature oil reservoirs across North Africa and deliver critical scientific insights to advance Egypt’s decarbonisation and climate-action goals under the National Climate Change Strategy 2050.

Abstract Wintertime explosive cyclones (ECs) pose severe wind-related hazards in the midlatitudes, yet the variability of their wind destructive potential (WDP) over the North Pacific and the dynamics driving this variability remain elusive. This study investigates the leading modes in WDP variability of wintertime North Pacific ECs and their formation processes. Two leading modes are identified by empirical orthogonal function analysis: EOF1 exhibits a monopole over the central North Pacific, whereas EOF2 displays a southwest–northeast dipole. These modes are dynamically linked to the Pacific/North American (PNA; EOF1) and East Pacific/North Pacific (EP/NP; EOF2) teleconnection patterns. The associated large-scale regimes—featuring a zonally extended upper-level jet in EOF1 and a northward-curving jet exit in EOF2—modulate baroclinicity and moisture convergence. Diagnostics of high-frequency available potential energy (APE H ) and kinetic energy (KE H ) during the EC lifecycle show that WDP anomalies form through two pathways. One is a local pathway, in which lower-tropospheric KE H is supplied by local mid-tropospheric baroclinic conversion of APE H generated through baroclinic and diabatic processes. The other is a downstream-development pathway, in which upper-tropospheric KE H dispersed downstream along the anomalous jet contributes to lower-tropospheric KE H . For positive WDP anomalies in both modes, the two pathways operate during cyclogenesis and rapid intensification, whereas the local pathway dominates at the lifetime maximum wind stage when the anomalies reach the poleward flank of the anomalous upper-level jet. In contrast, the negative WDP anomaly over the northwestern Pacific in EOF2 arises primarily from reduced diabatic APE H generation and weaker local baroclinic conversion.