Southampton Solent University

An extreme rainfall event that occurred from 16 to 18 December 2021 along the coastal regions of Peninsular Malaysia (PM) caused widespread flooding and substantial socioeconomic impacts. This study investigates the mechanisms leading to this event, focusing on the roles of climatic phenomena and local terrains. Two atmospheric interactions play key roles in triggering the event. Firstly, a strong cold surge (CS) associated with the East Asian winter monsoon (EAWM) interacted with the easterly surge over the southern South China Sea, leading to the formation of Borneo vortex. Secondly, a strong northeasterly and CS largely contributed to enhancing and transporting the vortex towards the PM and across the Titiwangsa mountain ranges. The phase change of the Indian Ocean Dipole (IOD) facilitated the eastward propagation of the vortex. Sumatra and PM terrains significantly modulated vortex evolution and moisture convergence over the Strait of Malacca. These findings are analyzed to shed light on interactions between large-scale climate drivers and localized terrain in generating extreme rainfall, emphasizing the necessity of multi-scale analysis for model accuracy.

Tuberculosis (TB) remains a major global health challenge, causing approximately 1.4 million deaths annually. In many high-burden regions, limited access to expert radiological interpretation leads to delayed or missed diagnoses. To address this, we propose a cost-effective, automated TB screening method suitable for under-resourced settings. Our method integrates a Convolutional Autoencoder Neural Network and a Multi-Scale Convolutional Neural Network with deep layer aggregation into an ensemble learning architecture for robust TB detection from chest radiographs. The framework was evaluated on two public datasets and one private dataset, achieving 99% sensitivity and 94% specificity on the Shenzhen dataset, and consistently high accuracy across all datasets. Expert radiologists reviewed a subset of the predictions, confirming the clinical relevance and diagnostic reliability of the model. The ensemble approach demonstrated strong generalisability, effectively identifying active pulmonary TB in chest X-rays from a globally representative cohort. It also outperformed existing classifiers, achieving a state-of-the-art Area Under the Receiver Operating Characteristic of 0.98. These results highlight the potential of our approach as a practical and scalable tool for TB screening, particularly in low- and middle-income countries where radiological resources are limited.

Global food security faces escalating threats from climate variability and resource constraints. Accurate crop yield forecasting is essential; however, existing methods frequently overlook complex spatial dependencies driven by climate teleconnections, such as the ENSO, and lacks rigorous uncertainty quantification. This paper presents HSE-GNN-CP, a novel framework integrating heterogeneous stacked ensembles, graph neural networks (GNNs), and conformal prediction (CP). Domain-specific features are engineered, including growing degree days and climate suitability scores, and explicitly model spatial patterns via rainfall correlation graphs. The ensemble combines random forest and gradient boosting learners with bootstrap aggregation, while GNNs encode inter-regional climate dependencies. Conformalized quantile regression ensures statistically valid prediction intervals. Evaluated on a global dataset spanning 15 countries and six major crops from 1990 to 2023, the framework achieves an R2 of 0.9594 and an RMSE of 4882 hg/ha. Crucially, it delivers calibrated 80% prediction intervals with 80.72% empirical coverage, significantly outperforming uncalibrated baselines at 40.03%. SHAP analysis identifies crop type and rainfall as dominant predictors, while the integrated drought classifier achieves perfect accuracy. These contributions advance agricultural AI by merging robust ensemble learning with explicit teleconnection modeling and trustworthy uncertainty quantification.

Purpose This study explores the impact of the shift to online education on relational pedagogy, particularly during the pandemic and the subsequent adaptation of hybrid learning. It aims to understand how online teaching affects the enactment of relational pedagogies from the lecturers' perspectives. Design/methodology/approach A duoethnographic method was employed, reflecting on the authors' experiences over three academic terms during the pandemic. Data were collected through five unstructured, conversational interviews between the two authors, focusing on their experiences and reflections. The analysis was conducted using Goffman's dramaturgical framework, applying theatrical metaphors to illustrate the performative dimensions of online teaching. Findings The transition to online education transformed relational pedagogical approaches, shifting teaching towards content-focused delivery. The absence of micro-interactions altered connections and reduced shared, co-productive learning. Lecturers compensated by delivering monologues aligned with curricula. Despite some positive outcomes, technological and disembodied complicated the enactment of relational pedagogy and in our view inhibited teaching effectiveness and performance repertoire. Originality/value This study offers a novel application of Goffman's theatrical metaphor to online pedagogy, revealing how the shift to digital spaces reconfigures the lecturer's role and relational engagement, specifically linked to micro-interactions. It offers recommendations for improving online teaching practices to better support positive relational pedagogies.

This critical review interrogates how contemporary diversity, equity, and inclusion (DEI) reforms in STEM education engage the deeper project of epistemic decolonisation. Framed by critical race theory, feminist science studies, and decolonial scholarship, it asks whether inclusion agendas move beyond representational expansion to disrupt Eurocentric hierarchies of legitimacy; which pedagogical and curricular innovations enact pluriversal STEM; and what institutional conditions constrain transformation. A multi-stage search of Scopus, Web of Science, ERIC, Google Scholar, and grey literature (2010–2025) yielded 152 records; PRISMA-informed screening produced 80 sources for interpretive thematic synthesis. Findings show that DEI initiatives have increased access and participation, yet typically preserve assumptions of scientific neutrality and universalism, leaving epistemic injustice largely intact. In contrast, decolonial innovations, such as two-eyed seeing, culturally sustaining and place-based pedagogies, history, philosophy, and sociology of science integration, and project-based learning grounded in indigenous knowledge systems, reposition learners and communities as co-producers of knowledge and reframe science as situated and relational. However, these practices remain peripheral due to assessment regimes, accreditation pressures, funding and tenure incentives, disciplinary gatekeeping, and limited educator preparation. The review argues that meaningful reform requires structural reconfiguration of curricula, evaluation, and institutional reward systems to recognise multiple epistemologies, cultivate ethical relationality, and enable sustained community partnership.