Atmospheric and Oceanic Sciences

Abstract As the dominant atmospheric circulation pattern over the western Tibetan Plateau (TP), the Western Tibetan Vortex (WTV) exerts substantial control on springtime 2 m surface air temperature ( T 2m ). However, its underlying radiative processes remain unclear. This study integrates GEWEX satellite observations with ERA5 and MERRA‐2 reanalysis, applying surface energy balance diagnostics to quantify the WTV's radiative forcing on T 2m variability. We find the WTV explains ∼66% of T 2m variance ( R = 0.81) across the western TP and the adjacent Southwest Asia. Downward shortwave radiation (DSW) emerges as the primarily radiative factor modulated by the WTV via cloud radiative forcing (CRF) processes. Specifically, anticyclonic WTV events reduce cloudiness, generating positive CRF anomalies that enhancing surface DSW and cause warming. Conversely, cyclonic events increase cloudiness, producing negative CRF anomalies that diminish DSW and induce cooling. These findings advance understanding of the radiative processes by which the upper circulations modulate the surface climate over the TP.

Abstract Controlling fluid injection is widely considered a key to limiting the size of injection‐induced seismicity, yet whether and how it limits earthquake size remains debated. We perform injection‐reactivation experiments on critically stressed faults to test how different injection strategies shape slip and seismic moment release. We find that injection strategies regulate the cadence of slip events rather than the total seismic moment. Compared to continuous injection, cycled injection triggers frequent and smaller events, reducing maximum moment magnitude and deformation energy of individual events. Injection–extraction cycles actively reduce pore pressure, temporally partition successive slip events, and effectively suppress delayed seismicity. Regardless of constant or cycled injection rates, maximum seismic moment () scales with cumulative injection volume () . Our laboratory results suggest that regulating the cadence of moment release promotes effective hazard mitigation.

Abstract This study uses a novel approach to infer the solar soft X‐ray (SXR) irradiance from Auger photoelectron fluxes. Solar SXR irradiance is highly variable, particularly during solar flares. Solar SXR photons at 2.5 nm produce oxygen Auger photoelectrons by photoionizing atmospheric constituents. More specifically, we derive the fit function between the Auger photoelectron fluxes observed by Mars Global Surveyor (MGS) and the 0.1–0.8‐nm solar irradiance observed by the Geostationary Operational Environmental Satellite (GOES) X‐ray Sensors (XRS). The fit function is applied to the MGS data to produce a proxy for the 0.1–0.8‐nm SXR irradiance and to classify 167 identified solar flare events from 1999 to 2006 at Mars, which is an additional vantage point away from Earth. These results are of particular interest for studying Mars's atmospheric response to solar flares and over‐the‐limb and far‐side solar activities.

Abstract The scarcity of ambipolar diffusion observations has constrained our understanding of ionospheric physical processes. We present the first comprehensive analysis of topside ionospheric ambipolar diffusion using the difference between the field‐aligned ion drifts and neutral winds observed by the Ionospheric Connection Explorer (ICON) satellite. The ambipolar diffusion exhibits a distinct evening enhancement within ±10° MLAT. The reversal of its direction near ±10° MLAT demarcates the boundary between the trans‐equatorial diffusion‐dominated region (within ±10° MLAT) and the area governed by the vertical gradient‐driven diffusion‐dominated region (out of ±10° MLAT). For the field‐aligned drift, the neutral wind contribution is also significant, but less influential than diffusion. The equatorial fountain effect further modulates diffusion, enhancing winter‐hemisphere transport and creating its longitudinal structures. These findings provide crucial observational evidence to strengthen the understanding of ionospheric diffusion mechanisms, offering reliable and critical inputs for ionospheric theoretical models.

Abstract Observational estimates of gravity wave momentum fluxes obtained from analyses of super‐pressure balloon tracks show an approximate log normal distribution in space and time. One study has suggested that a non‐orographic gravity wave parameterization with a steady source could reproduce this distribution through variability in the background wind, which “dynamically filters” the source spectrum. We use an implementation of a simple steady source gravity wave parameterization to show that this observed distribution is not reproduced, the most prominent departure from it being a large negative skew. We perform a crude tuning of the scheme's parameters to show that the discrepancy in the distribution cannot be corrected. We observe that at higher levels, parameterized fluxes are still skewed but less so. We suggest that dynamical filtering alone cannot introduce enough variability to obtain fluxes that are log normally distributed.

Abstract Understanding how the short‐term evolution of synoptic weather patterns influence Mesoscale Convective Systems (MCSs) is essential, as these systems are responsible for over half of central U.S. flash floods, leading to substantial socioeconomic and water resource management impacts. This study analyzes long‐term MCS data, flash flood reports, and atmospheric reanalyses from 2007 to 2017 using a machine learning clustering algorithm to examine how the synoptic weather patterns evolve prior to MCS initiation. While the clusters reflect seasonal and regional differences in MCS occurrence, they do not consistently distinguish between MCSs that do and do not produce flash floods. Systems in the southern Great Plains are more flood‐prone when a synoptic‐scale forcing, located near the system, drives strong water vapor transport from the nearby moisture source. More generally under different synoptic weather patterns, a broader precipitating area is the most dominant factor governing MCS flash flood potential.

Abstract Enabled by state‐of‐the‐art electric field measurements from the Van Allen Probes and careful calibration of the high‐quality data, we developed the first machine‐learning based inner‐magnetosphere electric field model, which covers L = 2.5–6.0 within 20 around the magnetic equator. The model output is the DC electric field perpendicular to the background magnetic field, including the poloidal and toroidal components. The most informative drivers, including the SYM‐H and AE indices and solar wind speed, are automatically identified during the model training process. The model input consists of these parameters along with time and spatial coordinates. The model successfully reproduces the storm‐time evolution of meso‐scale electric field structures, potentially related to subauroral polarization streams and dawnside auroral polarization streams. Given the growing recognition that meso‐scale electric field structures modulate the transport of high‐energy electrons, our model can incorporate these structures into studies of ring current and radiation belt dynamics.

Abstract Salt precipitation has emerged as a critical factor affecting injectivity, reservoir stability, and the potential to trigger near‐wellbore microseismic activity during geological CO 2 sequestration. While previous studies have primarily focused on the brine acidification induced by CO 2 injection, triggering geochemical reactions in carbonate rocks and leading to mechanical degradation, the mechanical behavior associated with salt precipitation in drying zones, particularly the failure mechanisms, remains poorly understood. In this work, we designed a reservoir‐condition displacement system to mimic near‐wellbore drying process and further investigated the rock failure modes due to salt precipitation in red sandstone samples. Our study demonstrates that, despite the densification of the pore structure due to salt precipitation, the overall mechanical performance of the rock undergoes significant deterioration. More importantly, for the first time, we observe a distinct transition of failure mode from shear‐to tensile‐dominated under uniaxial compression. Microstructural analysis further shows that the growth of polycrystalline and bulk crystals induces microcrack initiation and propagation, with the failure mechanism of rocks subjected to salt precipitation primarily characterized by intercrystalline damage at weak bonding interfaces under external loading.

Abstract Increase in pre‐monsoon heatwaves over the Indo‐Gangetic Plains (IGP) of India threatens life and ecosystems. We investigate the sufficiency of large‐scale anticyclones in forming moist and dry heatwaves by analysing local land‐atmospheric conditions prior to 10 heatwaves in IGP. Using Eulerian temperature decomposition, we show that horizontal heat advection (remote) process contributes minimally to both heatwave types compared to diabatic and adiabatic (local) processes. Antecedent factors highlight the critical role of pre‐monsoon showers, followed by nocturnal low‐level clouds and surface moisture, creating a conducive environment for moist heatwaves. In contrast, dry heatwaves form due to anticyclonic conditions, high sensible heat flux, lack of moisture advection, and cloud‐free conditions. This study highlights the role of local land‐atmosphere interactions in triggering the onset of heatwaves and help in distinguishing heatwave regions from adjacent non‐heatwave regions. Monitoring these precursors is essential for developing reliable early warning systems to improve heatwave detection and response.

Abstract North China has recently seen frequent hot days above 35°C or even 40°C before or at the very beginning of June. This raises a concern about the changing seasonality of exceptional hot extremes in the region which might leave population there underprepared. Based on station observations, we found that since 1990 early summer hot extremes have warmed 2–3 times faster than those in peak and late summer stages. A Lagrangian decomposition points to the main driver for the intra‐seasonally disproportionate warming rates as the enhanced adiabatic heating, contributed by more frequent arrival and intensified descent of upper‐level air particles from west and south. We further illustrated that the past‐decade prevalence of a ridge pattern well atop North China was essential to the enhanced adiabatic heating. Our results highlight the critical yet overlooked role of changing atmospheric dynamics in altering regional extremes.

Deafness poses significant challenges to effective communication, particularly in contexts where access to sign language interpreters is limited. Hand configuration recognition represents a fundamental component of sign language understanding, as configurations constitute a core cheremic element in many sign languages, including Italian Sign Language (LIS). In this work, we address configuration-level recognition as an independent classification task and propose a machine vision framework based on RGB-D sensing. The proposed approach combines MediaPipe-based hand landmark extraction with normalized three-dimensional geometric features and a Support Vector Machine classifier. The first contribution of this study is the formulation of LIS hand configuration recognition as a standalone, configuration-level problem, decoupled from temporal gesture modeling. The second contribution is the integration of sensor-acquired RGB-D depth measurements into the landmark-based feature representation, enabling a direct comparison with estimated depth obtained from monocular data. The third contribution consists of a systematic experimental evaluation on two LIS configuration sets (6 and 16 classes), demonstrating that the use of real depth significantly improves classification performance and class separability, particularly for geometrically similar configurations. The results highlight the critical role of depth quality in configuration-level recognition and provide insights into the design of robust vision-based systems for LIS analysis.

Recent advances in deep learning (DL) have enabled the integration of diverse biomedical data for disease prediction and risk stratification. Building on this progress, the overall objective of this study was to develop and evaluate a multimodal DL framework for robust multi-label classification (MLC) of major comorbidities in patients with obstructive sleep apnea (OSA) using physiological time series signals and clinical data. This study proposes a robust framework for multi-label classification (MLC) of comorbidities in patients with OSA using diverse physiological and clinical data sources. We conducted a retrospective observational study including a convenience sample of 144 patients referred for overnight polysomnography at the Sleep Medicine Center (SleepLab Split), University Hospital Centre Split (KBC Split), Split, Croatia. Patients were selected based on predefined inclusion criteria and data availability. A one-dimensional Convolutional Neural Network (1D-CNN) was developed to process and fuse time series signals, oxygen saturation (SpO2), derived SpO2 features, and nasal airflow (FP0), with demographic and physiological parameters, enabling the identification of key comorbidities such as arterial hypertension, diabetes mellitus, and asthma/COPD. The instruments included polysomnography-derived signals (SpO2 and FP0 airflow) and structured demographic/physiological parameters. Signals were preprocessed and used as inputs to the proposed fusion model. The proposed model was trained and fine-tuned using the Optuna hyperparameter optimization framework, addressing class imbalance through weighted loss adjustments. Its performance was comprehensively assessed using multi-label evaluation metrics, including macro/micro F1-score, AUC-ROC, AUC-PR, subset and partial accuracy, Hamming loss, and multi-label confusion matrix (MLCM). The study protocol was approved by the Ethics Committee of the School of Medicine, University of Split (Approval No. 003-08/23-03/0015, Date: 17 October 2023). The 1D-CNN achieved superior predictive performance compared to traditional machine learning (ML) classifiers with macro AUC-ROC = 0.731 and AUC-PR = 0.750. The model demonstrated consistent behavior across age, gender, and BMI groups, indicating strong generalization and minimal demographic bias. In conclusion, the results confirm that SpO2 and airflow signals inherently encode comorbidity-specific physiological patterns, enabling efficient and scalable screening of OSA-related comorbidities without the need for full polysomnography. Although the study is limited by data set size, it provides a methodological basis for the application of multi-label DL models in clinical decision support systems. Future research should focus on the expansion of multi-center datasets, thereby improving model interpretability and potential clinical adoption.

In recent years, the methods based on convolutional neural networks have achieved significant progress in hyperspectral image super-resolution. However, existing methods still face two key challenges: (1) they fail to fully extract edge detail information from hyperspectral images; (2) they struggle to simultaneously capture local and global features. To address these issues, we propose an Edge-Distilled and Local–Global Feature Selection network (EDLGFS) for hyperspectral image super-resolution. This network aims to effectively leverage edge details and local–global features, thereby enhancing super-resolution reconstruction quality. Firstly, we design an edge-guided super-resolution network based on knowledge distillation. This network transfers edge knowledge to improve the reconstruction. Secondly, we propose a Local–Global Feature Selection mechanism (LGFS), which integrates convolutions of different sizes with the self-attention mechanism. This design models spatial correlations across features with different receptive fields, achieving efficient feature selection to more effectively capture local and global features. Finally, we propose a dynamic loss mechanism to more effectively balance the contribution of each loss term. Extensive experimental results on three public datasets demonstrate that the proposed EDLGFS achieves superior super-resolution reconstruction quality.

The Global Navigation Satellite System (GNSS) is commonly used for outdoor positioning. However, its effectiveness diminishes in urban canyons and indoor environments attributed to signal blockage. This study aims to explore the potential of GNSS signals penetrating indoor spaces through windows and to enhance indoor positioning with 3D Mapping-Aided (3DMA) GNSS, a concept generally applied outdoors. The research employs a 3D model of a corridor with manually labeled window locations to predict satellite visibility within indoor areas. The study integrates Pedestrian Dead Reckoning (PDR) with an indoor Shadow-matching (I-SM) technique, utilizing an Extended Kalman Filter (EKF) to improve positioning accuracy. One of the findings indicates that the proposed method significantly enhances positioning performance and its availability, achieving a root mean square error (RMSE) that is 2 m better than using PDR alone or single epoch I-SM. The study concludes that integrating GNSS with I-SM technique and PDR can optimize an indoor positioning solution and highlights the potential for improved navigation solutions in complex urban environments.