Ocean radiocarbon (14C) is a proxy for air-sea exchange, vertical and horizontal mixing, and water mass identification. Here, we present five pre- to post-bomb coral Δ14C records from West Flower Garden Bank and Santiaguillo reefs in the Gulf of Mexico, Boca de Medio, and Isla Tortuga near the Cariaco Basin north of Venezuela. To assess basin-wide Δ14C variability, we compiled the Atlantic Ocean reef-building surface coral Δ14C records (24 corals and 28 data sets in total) with these new records. Cumulatively, the Δ14C records, on their independent age models, reveal the onset of post-bomb Δ14C trends in 1958 ±1 to 2 years. A general decrease in maximum Δ14C values occurs with decreasing latitude reflecting the balance between air-sea gas exchange and surface water residence time, vertical mixing, and horizontal advection. A slightly larger atmospheric imprint in the northern sites and relatively greater vertical mixing and/or advection of low-14C waters influence the southern Caribbean and eastern Atlantic sites. The eastern Atlantic sites, due to upwelling, have the lowest post-bomb Δ14C values. Equatorial currents from the eastern Atlantic transport low Δ14C water towards the western South Atlantic and southern Caribbean sites. Decadal Δ14C averages for the pre-bomb interval (1750–1949) for the low latitude western Atlantic are relatively constant within analytical (3–5‰) and chronological uncertainties (∼1–2 years) due to mixing and air-sea exchange. The compiled Δ14C records provide updated regional marine Δ14C values for marine reservoir corrections.

The Cunene region of southern Angola, especially the area around Xangongo, hosts a large number of African baobabs, including several superlative specimens. Our research reports the investigation of the three biggest specimens from Xangongo Grove, named XG-1 (11 stems; circumference 26.34 m), XG-2 (5 stems; 22.70 m) and XG-3 (9 stems; 27.73 m), and of the largest baobab from Xangongo town, named XT-1 (7 stems; 21.02 m). Several wood samples were collected from these four baobabs mainly as cores from trunks and/or primary branches and then radiocarbon-dated by AMS. The two oldest samples were extracted from two primary branches of baobab XG-1. These have practically identical radiocarbon dates of 1822 ± 19 BP and 1822 ± 10 BP, which correspond to identical calibrated ages of 1785 ± 15 and 1785 ± 10 calendar years. The calculated age of baobab XG-1 is 2100 ± 50 years. Thus, XG-1, which is called by the locals “The biggest baobab of Africa”, becomes the oldest living African baobab with accurate dating results. The results indicate ages of 1100 years for baobab XG-2, 850 years for baobab XG-3 and 550 years for baobab XT-1.

We report a set of radiocarbon dating of prehistoric settlements located on the paleolake Baruun Khuree shores in the Gobi-Altai area, southern Mongolia. The obtained series of 11 AMS 14C measurements on charcoal and other charred plant macro-remains can be associated with one of the earliest episodes of the Holocene highly mobile desert-adapted hunter-gatherers activities from the Gobi desert (ca. 11,250–10,500 cal BP). Exploiting a wide range of environments, including dune fields, they are characterized by pottery usage and microblade core technology with wedge-shaped cores as well as osteological materials. These preliminary results are part of a project analyzing the nature of long-lasting prehistoric occupation around Tsakhiurtyn Hundi (Eng. Flint Valley)—one of the most extensive early prehistoric sites of Central Asia owing its name to the presence of abundant flint outcrops, lithic workshops and their innumerable flint artifacts.

Machine-learning (ML) methods have shown great potential for weather downscaling. These data-driven approaches provide a more efficient alternative for producing high-resolution weather datasets and forecasts compared to physics-based numerical simulations. Neural operators, which learn solution operators for a family of partial differential equations, have shown great success in scientific ML applications involving physics-driven datasets. Neural operators are grid-resolution-invariant and are often evaluated on higher grid resolutions than they are trained on, i.e., zero-shot super-resolution. Given their promising zero-shot super-resolution performance on dynamical systems emulation, we present a critical investigation of their zero-shot weather downscaling capabilities, which is when models are tasked with producing high-resolution outputs using higher upsampling factors than are seen during training. To this end, we create two realistic downscaling experiments with challenging upsampling factors (e.g., 8x and 15x) across data from different simulations: the European Centre for Medium-Range Weather Forecasts Reanalysis version 5 (ERA5) and the Wind Integration National Dataset Toolkit. While neural operator-based downscaling models perform better than interpolation and a simple convolutional baseline, we show the surprising performance of an approach that combines a powerful transformer-based model with parameter-free interpolation at zero-shot weather downscaling. We find that this Swin-Transformer-based approach mostly outperforms models with neural operator layers in terms of average error metrics, whereas an Enhanced Super-Resolution Generative Adversarial Network-based approach is better than most models in terms of capturing the physics of the ground truth data. We suggest their use in future work as strong baselines.