2385 papers
This study reports the discovery of a third *Megantereon* form, *M. microta*, in the Early Pleistocene Longdan Fauna of Northwestern China, which, alongside other Oriental species, provides key evidence for a humid ecotone facilitating faunal interchange between northern and southern provinces during that period.
This paper proposes an interface-aware method for safe aggregation in modular sustainability assessment that uses a factorization criterion to identify the coarsest intermediate flow representations, thereby enabling valid local Pareto screening that preserves system-level optima and avoids the false elimination of alternatives caused by premature nominal-product pruning.
This paper introduces a global deep learning model trained on 760 tide gauges that successfully forecasts storm surges in unseen locations with significantly higher accuracy than persistence and global hydrodynamic reanalysis, demonstrating that forecast skill scales with the diversity and number of training gauges.
This paper introduces the Physics-Informed Cross-Modal Attention (PI-CMA) model, which aligns geomagnetic, seismic, and hydrologic data to reveal that while the method achieves competitive predictive performance, its learned attention weights primarily reflect baseline structural correlations rather than statistically significant, event-specific earthquake precursors.
This study demonstrates that ocean acidification, particularly when assessed via winter minimum aragonite saturation states rather than annual averages, is already pushing a significant portion of productive marine regions below critical thresholds for calcifying organisms, thereby compounding existing anthropogenic stressors and necessitating immediate updates to monitoring protocols and risk assessments.
The study reveals that declining summer sea ice exposes Antarctic ice shelves to unprecedented wave erosion by warm surface waters, driving rapid frontal retreat rates that are currently omitted from sea-level projections and threaten the stability of nearly all Antarctic ice shelves.
This study demonstrates that integrating continuous multi-parameter sensor networks with periodic Terrestrial Laser Scanning surveys effectively quantifies volumetric changes and captures precursor micro-deformations in the soft limestone cliffs of the Bonifacio Citadel, thereby providing critical decision-support data for managing rockfall hazards at vulnerable coastal heritage sites.
This study utilizes ANSYS Fluent to validate a numerical model and demonstrates that in full-scale free surface constructed wetlands, strategic vegetation distribution patterns (specifically banded and fully vegetated configurations) significantly enhance hydraulic performance and flow uniformity, proving more critical to efficiency than planting density alone.
This study utilizes a reproducible Google Earth Engine workflow with Sentinel-1 SAR imagery to map flood extents in Punjab, Pakistan from 2015 to 2025 and integrates these spatial findings with participatory assessments to inform livestock disaster preparedness and resilience strategies.
This study utilizes a Cross-Survey Magnetic Persistence Framework to analyze heterogeneous aeromagnetic data from Uruana de Minas, Brazil, revealing that raw 2011 observations demonstrate a localized magnetic-gradient anomaly characterized by a northward gradient reversal that persists across multiple surveys and processing states.