Tag: Atmospheric Science
Self-supervised graph model improved multi-horizon weather forecasts
What the study found The study found that a self-supervised spatio-temporal graph model improved multi-variable weather forecasting across multiple forecast horizons. The authors report that it performed better than traditional numerical weather prediction models and recent deep learning methods on the datasets they tested. Why the authors say this matters The authors conclude that the…
Synthetic bed choices affect Antarctic ice-loss projections
What the study found Small differences in synthetic bed topography can lead to noticeable differences in projected Antarctic ice mass loss and in the timing and extent of grounding line retreat. The study found that sea-level rise estimates varied depending on how the synthetic bed was generated and whether basal friction coefficients were optimized for…
Bed topography strongly affects Thwaites Glacier mass loss
What the study found The specific bed topography beneath Thwaites Glacier had a first-order control on accumulated mass loss. Final sea-level rise did not scale with bed resolution. Why the authors say this matters The authors conclude that continued high-resolution topography mapping is important. The findings indicate that current projections may underestimate uncertainty linked to…
Automating future Antarctic bed maps remains challenging
What the study found Future gridded Antarctic ice-sheet and bed mapping could potentially be automated to speed up the delivery of new datasets. The article identifies several recurring problems that make this difficult, including survey disagreements, large data gaps, changing ice thickness, and interpolation methods that do not work equally well across landscapes. Why the…
Machine learning can improve ice sheet bed mapping
What the study found The study found that machine learning can enhance analysis of airborne radio-echo sounding data used to map ice-sheet bed topography. The authors highlight uses in denoising radar data, automatically picking radar returns, and improving spatial interpolation and uncertainty quantification. Why the authors say this matters The authors say this matters because…
New BedMachine updates improve Antarctic and Greenland bed maps
What the study found The study reports major improvements in the description of the bed topography and bathymetry, meaning the shape of the land under the ice and the ocean floor, for both the Antarctic and Greenland Ice Sheets. Why the authors say this matters The authors say these improvements address known limitations in BedMachine,…
Bed topography uncertainty strongly affects Antarctic ice-sheet projections
What the study found Uncertainty in bed topography, the shape and elevation of the land beneath the Antarctic Ice Sheet, can substantially affect simulations of Antarctica's future evolution. The abstract reports that these errors can change the Antarctic contribution to sea level by more than 40 cm by 2150 and 1 m by 2300. Why…
Tropical isoprene variability differs across three regions
Study reveals tropical isoprene varies by region: Amazonia emission-controlled, Maritime Continent chemistry-controlled, and equatorial Africa intermediate, requiring region-specific atmospheric.
Neural network predicts shifts in extreme weather frequency
—
in MeteorologyNeural networks leverage climate model data to predict how extreme rainfall, hail, and winds will shift geographically as climate changes, accounting for terrain effects.
Landfalling South Pacific atmospheric rivers are projected to intensify
—
in MeteorologyStudy projects atmospheric river frequency to double over South Pacific by mid-century, with robust trends emerging within 10-20 years first affecting southern New Zealand and Tasmania.
