Search Results (2,839)

Climate change poses a significant threat to water resources and agricultural sustainability, particularly in semi-arid and socio-economically vulnerable regions such as South Africa. This review synthesizes empirical, modelling, and policy-based evidence on the impacts of climate change on water availability, crop production, and adaptation strategies in the country, drawing on approximately 162 peer-reviewed studies and institutional reports published between 2010 and 2025. The findings indicate that rising temperatures, shifting rainfall patterns, and an increasing frequency of extreme events, such as droughts and floods, are intensifying water stress and disrupting agricultural systems. Hydrological models consistently project declines in runoff, soil moisture, and streamflow, while crop simulation models predict reductions in the yields of major staple crops, including maize, wheat, and sorghum, particularly under high-emission scenarios. Although localized improvements in water availability and crop productivity may occur, these tend to be limited and highly context-specific. In response, South Africa has implemented a range of adaptation strategies, including climate-smart agriculture, water-efficient irrigation, ecosystem-based approaches, and policy-driven interventions. However, their effectiveness remains constrained by institutional fragmentation, limited financial capacity, and persistent socio-economic inequalities, particularly among smallholder farmers. The review underscores the need for integrated, inclusive, and context-specific adaptation strategies that strengthen governance, enhance the science–policy interface, and improve access to climate finance. The insights provided offer valuable guidance for advancing climate resilience in South Africa and other vulnerable regions across the Global South. Full article

Expansive clay slopes are vulnerable to progressive strength loss induced by repeated wetting and drying, a mechanism that drives shallow failure in active moisture zones. Reproducing this degradation experimentally is time-consuming and resource-intensive. This study evaluates whether Fully Softened Strength (FSS) can serve as a practical substitute for five wet–dry cycles in expansive clay slope stability assessment. Direct shear tests were conducted on wet–dry-cycled and reconstituted FSS specimens across fourteen experimental water contents. Strength parameters were incorporated into homogeneous and heterogeneous limit equilibrium slope models, considering degraded layer thicknesses of 1–5 m and suspended water table conditions. Equivalence was assessed using root mean square error (RMSE), prediction bias, and physical representativeness. Five wet–dry cycles produced a dominant cohesion reduction of 70.4% with minor changes in friction angle, reaching a quasi-stationary degraded state. FSS reproduced an equivalent system response through mechanical compensation between cohesion and friction—not through equality of strength parameters—under shallow failure conditions. The best statistical fit was obtained at w = 43.5% (RMSE = 0.314); however, w = 42.0%, coinciding with the liquid limit, provided a physically more robust interpretation with near-zero bias. Equivalence was found to be valid only for normal stresses ≤ 50 kPa, representative of shallow failure depths of 1–4 m. Full article

Multi-label image classification (MLIC) is vulnerable to contextual bias, where models may exploit spurious label–context associations rather than object evidence, leading to degraded generalization under distribution shifts. To address this issue, we propose CPAN, a causal-inspired framework that integrates label-specific feature decoupling, prototype-based mediator modeling, patch-level evidence aggregation, and adaptive fusion. Specifically, CPAN uses a Transformer decoder to extract label-specific representations from the whole image and local patches. We introduce a prototype dictionary as a surrogate mediator space to encourage the model to rely on object-relevant intermediate patterns rather than context-sensitive shortcuts. We further aggregate patch-level predictions to enhance direct object evidence and fuse them with whole-image predictions through a learnable gate. Experiments on two benchmark datasets show that CPAN consistently improves both recognition accuracy and robustness. On MS-COCO, CPAN achieves 85.26 mAP, 80.67 CF1, and 82.52 OF1; on NUS-WIDE, it reaches 66.11 mAP, 64.42 CF1, and 75.95 OF1. Under context-shifted evaluation on MS-COCO, CPAN further obtains 80.93 mAP, 75.84 CF1, and 77.87 OF1, indicating stronger robustness to contextual bias. These results show that CPAN learns more object-centered representations and reduces reliance on spurious contextual correlations. Full article

Sustainable groundwater management represents a main goal for the future in the context of climate change and increasing anthropogenic pressure. In recent decades, intrinsic vulnerability assessment and risk mapping have been established as some of the most important tools for groundwater preservation, but they have also shown limitations due to their static nature and their failure to account for the inherent uncertainty of hydrogeological parameters. This study proposes an innovative hybrid framework that integrates traditional overlay-index methodology (SINTACS Release 5) with stochastic numerical modeling to assess groundwater contamination risk and evolve it into a dynamic time-dependent tool. This methodology was applied to a case study of the Lapisina Valley phreatic aquifer (Northeastern Italy), a strategic area for drinking water supply. Numerical simulations were implemented to reproduce groundwater flow using the MODFLOW-NWT code. To address parametric uncertainty, 237 stochastic realizations of the modeling domain were generated using the Latin Hypercube Sampling method, randomizing hydraulic conductivity values. Advective transport was simulated through forward particle tracking using the MODPATH code, starting from the identified and classified hazard sources within the study area. Assuming the absence of attenuation during transport allowed for a conservative worst-case scenario. The result was the definition of a probabilistic contaminant propagation factor, a time-dependent indicator that quantifies the probability of pollution arrival to a specific discrete portion of the domain. This probabilistic factor was combined with three indexes commonly utilized for risk assessment (the intrinsic vulnerability index, hazard index, and value of the resource) to generate four contamination risk maps representing different timestep scenarios (5, 10, 20, and 50 years) after the arrival of a hypothetical contaminant in the saturated zone. This approach transforms risk mapping from being a useful but static snapshot to a predictive dynamic framework. Full article

Artificial intelligence (AI) is increasingly used in cancer research, enabling integrative analysis of complex biomedical data to identify actionable therapeutic vulnerabilities. This review specifically examines how AI advances mechanistic cancer target discovery and translational drug development, focusing on: (1) the processing of large-scale genomics, transcriptomics, proteomics, metabolomics, single-cell profiling, spatial, and clinical datasets using machine learning (ML) and deep learning (DL) algorithms; (2) the identification of candidate biomarkers, driver genes, dysregulated pathways, tumor dependencies, and molecular targets that traditional methods often miss; (3) the integration of multi-omics data, network biology, causal inference, and systems-level modeling to refine mechanistic understanding of cancer progression and separate functional driver events from passengers; and (4) applications in drug development, including virtual screening, molecular modeling, structure-informed target validation, drug repurposing, synthetic lethality prediction, and de novo drug design, which collectively may enhance early-stage drug discovery efficiency. The review underscores that AI serves as both a predictive tool and a platform for linking molecular mechanisms to hypothesis generation, target prioritization, and rational treatment design. Challenges such as data heterogeneity, algorithmic bias, interpretability, reproducibility, regulatory requirements, and patient privacy must be addressed for robust translation and clinical use. Future directions may focus on hybrid approaches that integrate causal modeling, explainable AI, multimodal data, and experimental validation to yield mechanistically grounded, clinically actionable insights. AI-driven approaches ultimately aim to accelerate mechanism-based cancer target discovery and enable more precise, biologically informed anticancer therapies. Full article

Background: Metabolic bariatric surgery (MBS) yields significant but heterogeneous recovery patterns. The longitudinal interplay between evolving eating behaviors and health-related quality of life (HRQoL) remains insufficiently characterized. Objectives: To identify trajectories of eating behavior and HRQoL during the first postoperative year and examine their associations with 12-month outcomes. Methods: A total of 244 patients from two hospitals in Taiwan were followed for 12 months. Dutch Eating Behavior Questionnaire, and Impact of Weight on Quality of Life-Lite were assessed. Group-based trajectory modeling (GBTM) identified latent subgroups, and multiple regression analyzed associations with 12-month HRQoL, adjusting for clinical covariates. Results: GBTM identified two distinct trajectories for restrained, emotional, and external eating. For HRQoL, three trajectories emerged: high-start stable (45–50%), moderate-decline (30–35%), and low-start improving (~20%). In the regression model (R² = 0.37, p < 0.001), eating behavior trajectories were not independently associated with total HRQoL at 12 months after adjusting for covariates, including baseline BMI and comorbidities. Specifically, restrained eating (β = −1.42, p = 0.502), emotional eating (β = −10.33, p = 0.110), and external eating (β = −5.33, p = 0.160) trajectories did not significantly predict global HRQoL scores. Conclusions: Postoperative adaptation is characterized by substantial heterogeneity, with a significant subgroup experiencing HRQoL decline despite surgery. While eating behavior trajectories align with domain-specific psychosocial trends, early postoperative clinical factors appear to exert a more dominant influence on total HRQoL during the first year. These findings suggest that multidisciplinary support should target specific vulnerable trajectories to optimize long-term outcomes. Full article

Background: Sleep is a critical recovery mechanism for athletes, supporting hormonal regulation, muscle repair, and cognitive function. Dual-career athletes are particularly vulnerable to sleep disruption, which may impair performance and stress regulation. This study examined the effects of a 14-day sleep hygiene intervention protocol (SHIP) on aerobic and anaerobic performance, as well as anticipatory cortisol responses, in university-level soccer players. Methods: Thirty athletes (females: n = 14, 22.1 ± 3.3 y, 157.8 ± 6.0 cm, 53.5 ± 3.9 kg, males: n = 16, 21.5 ± 1.7 y, 167.5 ± 5.9 cm, 62.7 ± 5.4 kg) completed the Pittsburgh Sleep Quality Index (PSQI), provided pre-test salivary cortisol samples, and performed the Yo-Yo Intermittent Recovery Test Level 1 (YYIR1) and Repeated Anaerobic Sprint Test (RAST) before and after the intervention (adhering daily to 10–18 individualized sleep hygiene). Results: The SHIP significantly reduced sleep latency (p = 0.04) and increased sleep duration (p = 0.03), and PSQI scores (p < 0.001) in both sexes. Females showed marked increases in sleep duration (p = 0.002), while males showed improved latency (p = 0.07). Five behaviourally coherent clusters derived from the SHIP adherence explained a substantial proportion of variance (74.99%). Stimulant and metabolic regulation, and bedroom light and thermal environment control consistently predicted sprint and physiological outcomes (p < 0.05). Anticipatory cortisol decreased before both tests (p = 0.03–0.04). YYIR1 performance improved for the full cohort (p = 0.001). RAST times slowed slightly (p = 0.02), though fatigue index improved (p = 0.05). Conclusions: A short-term SHIP effectively enhanced sleep, reduced physiological stress, and improved key performance outcomes in collegiate athletes. Full article

In the Mediterranean principality of Thessaly, Greece, heat stress has become an environmental limitation on animal production and welfare. This study aims to quantify livestock heat stress using the temperature–humidity index (THI) and assess its spatial and temporal distribution across Thessaly during the warm seasons from 2020 to 2025, based on ERA5-Land reanalysis data. For selected livestock units, hourly air temperatures and dew point temperatures were used to generate and calculate maximum temperature fields and the THI under outdoor conditions, with no directly measured physiological responses in animals, but potential heat stress exposure was evaluated using THI derived from ERA5-Land data. The results reveal persistent thermal hotspots in the central and southeastern Thessalian plain, where maximum daily temperatures frequently exceeded 38–40 °C and locally surpassed 45 °C during August. THI values regularly exceeded 72, indicating productivity decline, and reached 82 during peak summer months, corresponding to high and severe stress categories. Mountainous regions were consistently 6–10 °C cooler and exhibited lower THI levels. Thermally stressful conditions extended from May through September, indicating sustained seasonal exposure rather than isolated heatwave events. The spatial coincidence between intensive livestock production and high-THI zones suggests structural vulnerability under current climate conditions. These findings offer a spatially explicit assessment of climate-driven thermal risk and support the development of targeted mitigation strategies and climate-resilient livestock management in Mediterranean agricultural regions. They also offer a data-driven foundation for integration into emerging Digital Twin frameworks for predictive livestock management. Full article

Sleep disorders, such as insomnia, obstructive sleep apnea (OSA), narcolepsy, REM sleep behavior disorder, and circadian rhythm disturbances, represent a rapidly expanding global health burden that is strongly associated with cardiovascular, metabolic, neurological, and psychiatric diseases. Advancements in wearable sensing technologies and Internet of Medical Things (IoMT) infrastructures have expanded the possibilities for continuous, home-based sleep assessment beyond conventional polysomnography laboratories. These Sleep Health Internet of Things (S-HIoT) systems combine multimodal physiological sensing (EEG, ECG, SpO₂, respiratory effort and actigraphy) with wireless communication and cloud-based analytics for automated sleep-stage classification and disorder detection. Nonetheless, the digitization of sleep medicine brings about significant cybersecurity concerns. The constant transmission of sensitive biomedical information makes S-HIoT networks open to anomalous traffic flows, signal manipulation, replay attacks, spoofing, and data integrity violation. Existing studies mostly focus on analyzing physiological signals and network intrusion detection independently, resulting in a systemic vulnerability of cyber–physical sleep monitoring ecosystems. With the aim of addressing this empirical deficiency, this review integrates emerging advances (2022–2026) in the AI-assisted categorization of sleep phases and IoMT anomaly detector designs on the finer analysis of CNN, LSTM/BiLSTM, Transformer-based systems, and a component part of federated schemes and the lightweight, edge-deployable intruder assessor models available. The aim of this study is to uncover a gap in the literature: integrated architectures to trade off audiences of faithfulness of physiological modeling with communication-layer security. To counter it, we present a single framework to include CNN-based spatial feature extraction, Bidirectional Long Short-Term Memory (BiLSTM)-based temporal models and Random Forest-based ensemble classification using a dual task-learning approach. We propose a multi-objective optimization framework to jointly optimize the performance of sleep-stage prediction and that of network anomaly detection. Performance on publicly available datasets (Sleep-EDF and CICIoMT2024) confirms that hybrid integration can be tailored to achieve high accuracy [99.8% sleep staging; 98.6% anomaly detection] whilst being characterized by low inference latency (<45 ms), which is promising for feasibility in real-time deployment in view of targeting edge devices. This work presents a comprehensive framework for developing secure, intelligent, and clinically robust digital sleep health ecosystems by bridging chronobiological signal modeling with cybersecurity mechanisms. Furthermore, it highlights future research directions, including explainable AI, federated secure learning, adversarial robustness, and energy-aware edge optimization. Full article

Rapid urbanization and the rapid accumulation of multi-source and multi-temporal Earth observation data are creating an increasing demand for remote sensing models that can flexibly support fine-grained monitoring beyond fixed label taxonomies. Open-vocabulary remote sensing image semantic segmentation (OVRSIS) aims to segment text-specified categories beyond a fixed label space with vision–language foundation models. However, dense remote sensing scenes make pixel–text matching highly vulnerable to semantic confusion and misalignment, owing to extreme scale variation, thin structures, repetitive textures, and prompt sensitivity. To address these challenges, we propose UGDMoE, an uncertainty-guided mixture-of-experts framework for OVRSIS. First, we design a domain-specific MoE decoder with three geometrically specialized experts—for slender structures, mid-scale objects, and large-region context—routed by the alignment-risk cue $U_0$. Second, we introduce a lightweight prompt–response estimation strategy that quantifies prediction dispersion across semantically equivalent prompts to derive $U_0$ in an annotation-free manner. Third, we develop prompt ensemble-based likelihood calibration (PELC), which takes the shared alignment-risk cue $U_0$ as input to calibrate prompt-specific logits before refinement. Finally, we design a lightweight uncertainty-aware structure refinement module that, guided by $U_0$, selectively fuses early visual features with segmentation logits to restore boundary continuity and connectivity of thin structures. We conduct extensive experiments on eight OVRSIS benchmarks under cross-dataset evaluation protocols. Trained on DLRSD, it achieves 46.97 m-mIoU and 63.31 m-mACC, surpassing the strongest baseline by 0.76 and 0.62 points; trained on iSAID, it reaches 37.47 m-mIoU and 58.52 m-mACC, improving over the strongest competitor by 0.71 and 0.61 points. UGDMoE consistently achieves state-of-the-art performance and remains robust under training-source changes. Full article

Background: Reliable preoperative identification of carotid plaque instability remains challenging. Although duplex ultrasound allows early detection of carotid stenosis, it does not consistently predict plaque biological behavior. MicroRNAs (miRNAs) are small non-coding RNAs that regulate gene expression and have been implicated in atherosclerotic progression and plaque destabilization. The tissue-level expression of miRNAs in carotid plaques and their relationship with histological vulnerability remain incompletely defined. Methods: This exploratory, pilot, hypothesis-generating study included patients undergoing carotid endarterectomy for asymptomatic high-grade carotid stenosis (>75% NASCET). Plaque vulnerability was assessed using a multiparametric approach combining preoperative duplex ultrasound features (including Gray Scale Median, GSM), intraoperative macroscopic evaluation, and a validated histological scoring system; only plaques with concordant classification across all three modalities were retained for molecular analysis. Total RNA including small RNA was extracted from plaque tissue and miRNA expression was measured by qRT-PCR on a panel of 47 candidate miRNAs. Data were analyzed descriptively. Results: Twenty-eight patients were initially enrolled; after application of strict vulnerability criteria, five plaques (three unstable, two stable) were selected for miRNA profiling. Among the 47 miRNAs assayed, miR-122 and miR-197 showed a consistent descriptive trend toward higher expression in plaques classified as unstable; these plaques also displayed histological features of vulnerability (lipid-rich necrotic cores and inflammatory infiltrates). Given the extremely limited sample size, no inferential statistical comparisons or multiple-testing corrections were performed. Conclusions: In this small, tissue-based exploratory analysis, miR-122 and miR-197 were more highly expressed in plaques with histological features of instability. Due to the small sample size, the effect estimates are unstable, and the findings should be used solely to inform the design and power calculations of future studies. We outline the need of a clear, pragmatic validation pathway based on replication in independent, larger cohorts with standardized tissue handling and blinded assessment and parallel evaluation of circulating miRNA levels to assess noninvasive biomarker potential. Indeed, these findings are preliminary and strictly hypothesis-generating; validation in larger, prospectively collected cohorts and integration with circulating biomarkers and imaging data are required before clinical application. Full article

Background: High-grade serous ovarian carcinoma (HG-SOC) remains the most lethal gynecological malignancy, largely due to intrinsic or acquired resistance to platinum-based chemotherapy. Although large-scale sequencing studies have delineated the genomic landscape of HG-SOC, clinically actionable biomarkers predictive of platinum response and outcome are still lacking. This study aimed to identify genomic alterations associated with platinum sensitivity, resistance, or refractoriness, and to assess their prognostic relevance. Methods: Tumor DNA from 24 HG-SOC patients with optimal cytoreductive resection, classified as platinum-sensitive (n = 9), platinum-resistant (n = 8), or platinum-refractory (n = 7) underwent targeted next-generation sequencing of 409 cancer-associated genes. Somatic variants were filtered and classified for oncogenicity using established criteria incorporating predicted functional impact, REVEL scores, and population allele frequencies. Associations between mutational profiles, platinum response, and overall survival (OS) were evaluated using Kaplan–Meier and Cox regression analyses. Key findings were validated in the TCGA ovarian serous carcinoma (TCGA-OV) dataset using survival analyses. Results: A total of 1367 protein-altering somatic variants across 301 genes were identified. While TP53 mutations were ubiquitous, platinum-resistant and platinum-refractory tumors showed enrichment of pathogenic alterations affecting DNA repair, transcriptional regulation, epigenetic modification, and oncogenic signaling, including FANCA, ATF1, MAF, NCOA2, PIK3CA, and TET1. Mutations in these genes were associated with reduced overall survival in exploratory analyses (median 2.5–9 months vs. 27.5–45 months). Multivariate analysis identified FANCA and ATF1 as potential independent predictors in exploratory modeling. In the TCGA-OV cohort, patients harboring pathogenic variants in a multi-gene panel derived from this study (excluding BRCA1/2) exhibited significantly worse survival compared with both BRCA1/2-mutated cases and the overall cohort. Conclusions: This exploratory study identifies a set of genomic alterations converging on transcriptional and epigenetic regulation, DNA repair, and oncogenic signaling that are associated with platinum resistance and adverse prognosis in HG-SOC. Independent validation in TCGA supports the potential clinical relevance of this mutational signature. These findings warrant further validation in larger prospective cohorts and functional studies to clarify their role as biomarkers of aggressive disease and therapeutic vulnerability. Full article

Deep-learning-based vulnerability detection has made substantial progress, but two limitations remain prominent. Sequence-based methods linearize source code and thus weaken the explicit modeling of control-flow and data-flow dependencies. Graph-based methods preserve program structure, yet conventional graph neural networks still have difficulty capturing long-range interactions in large code property graphs (CPGs). In addition, standard CPGs usually lack explicit variable semantics and security-critical node roles, which limits their ability to represent vulnerability-relevant program behavior. To address these issues, we propose SV-GEN, a vulnerability detection framework that combines large-language-model-driven semantic enhancement with hybrid sequence-graph learning. The novelty of SV-GEN lies in introducing a semantically enriched code property graph, termed Sem-CPG, which augments conventional CPGs with variable semantic roles and security-oriented node labels, and in coupling this representation with an adaptive fusion mechanism over structural and sequential views. Specifically, we use a large language model as an external semantic annotator to assign variable roles and identify source, sink, and sanitizer nodes, and then encode the resulting Sem-CPG with a Graph Transformer while modeling the code sequence with GraphCodeBERT. A learnable gating module is further used to adaptively fuse the graph-level and sequence-level representations for final prediction. Experiments on Devign, ReVeal, and DiverseVul show that SV-GEN achieves competitive or superior overall performance across benchmarks, with particularly strong improvements on the large and highly imbalanced DiverseVul dataset. Full article

Background: Linezolid is widely used for the empirical and targeted treatment of Gram-positive infections. Elderly patients frequently exhibit substantial pharmacokinetic variability due to age-related physiological changes and high comorbidity burden, which may predispose to drug accumulation and toxicity. This study aimed to develop and evaluate a population pharmacokinetic (PopPK) model of intravenous (IV) linezolid in elderly patients (65–87 years) to support therapeutic drug monitoring and explore exposure-risk scenarios associated with overexposure. Methods: A retrospective single-center study including 149 patients and 293 serum trough concentrations was conducted. Patients were randomly assigned to development (n = 103) and independent validation (n = 46) cohorts. Linezolid concentrations were quantified using an enzyme immunoassay. The PopPK model was developed in NONMEM^® using FOCE-I. Model performance was evaluated using standard diagnostic plots, bootstrap analysis, visual predictive checks, and validation metrics (mean prediction error [MPE] and mean absolute prediction error [MAPE]). Monte Carlo simulations assessed the probability of overexposure (Cmin > 8 mg/L) and the probability of target attainment (PTA; AUC24/MIC ≥ 100) under standard dosing (600 mg IV every 12 h). Results: Linezolid pharmacokinetics were best described by a one-compartment model with first-order elimination. Estimated glomerular filtration rate, treatment duration, and age were identified as significant predictors of clearance. Internal and independent validation confirmed the robustness and predictive performance of the model. Simulations showed a high probability of overexposure in patients with impaired renal function, particularly during prolonged treatment. Conclusions: Renal function, age, and treatment duration are major determinants of linezolid exposure in elderly patients. Standard dosing frequently results in overexposure, supporting early therapeutic drug monitoring and individualized dose adjustment in this vulnerable population. Full article

Under climate change, extreme wind events are predicted to become both more common and more severe, increasing the vulnerability of plantation forests. In February 2023, ex-tropical Cyclone Gabrielle caused widespread wind damage to radiata pine (Pinus radiata D. Don) forests across the North Island of New Zealand, providing a rare opportunity to evaluate mechanistic wind-risk modelling under extreme storm conditions. This study assessed the performance of the ForestGALES model in predicting wind damage within the Rangipo genetic accelerator trial and examined the influence of stocking and cultivation on wind vulnerability. Using detailed pre-cyclone field measurements and high-resolution unmanned aerial vehicle light detection and ranging (ULS) data, ForestGALES was parameterised for the Rangipo trial and applied at both individual-tree and stand scales. Model predictions were compared with observed post-cyclone damage using balanced area under the receiver operating characteristic curve (AUC), accounting for strong class imbalance. Wind damage was observed in 16.7% of trees, of which 10.2% showed stem breakage and 6.5% overturning. Across both spatial scales, overturning was more accurately predicted than stem breakage. At the individual-tree scale, ForestGALES showed moderate predictive skill, with balanced AUC values of 0.650 ± 0.014 for overturning, 0.595 ± 0.011 for breakage, and 0.621 ± 0.008 for total damage. Model performance was stronger at the stand scale, where discrimination was highest for overturning (AUC 0.811 ± 0.122), followed by breakage (0.693 ± 0.116) and total damage (0.623 ± 0.083). Silvicultural treatments significantly influenced predicted critical wind speeds (CWS). High-stocking treatments exhibited consistently higher CWS values and therefore greater wind firmness than low-stocking treatments, while cultivation effects were smaller but significant. Simulated reductions in stocking further demonstrated increased wind vulnerability as stocking declined, highlighting thinning as a primary determinant of wind risk. These findings demonstrate that ForestGALES can reliably discriminate wind damage at operational stand scales under extreme cyclone conditions and highlight the importance of stand structure in improving plantation resilience under increasingly storm-prone climates. Full article