New research evaluates the performance of generative adversarial networks for stochastic parameterizations.
machine learning & AI
Combining AI and Analog Forecasting to Predict Extreme Weather
New deep learning technique brings an obsolete forecasting method “back to life” to predict extreme weather events.
Representing Estuaries and Braided Rivers as Channel Networks
The human eye is quite good at identifying channel networks among the rich patterns exhibited by estuaries and braided rivers, but computers have a harder time doing so. Could they do better?
New Study Hints at Bespoke Future of Lightning Forecasting
Researchers used machine learning to develop a model that can predict lightning strikes to within 30 minutes of their occurrence and within 30 kilometers of a weather station by using just four simple atmospheric measurements.
Using Satellites and Supercomputers to Track Arctic Volcanoes
New data sets from the ArcticDEM project help scientists track elevation changes from natural hazards like volcanoes and landslides before, during, and long after the events.
Using a Machine to Help Us Learn About Jupiter’s Aurora
A first usage of principal component analysis on Hubble images of Jupiter’s auroral ovals reveals the most common patterns, and machine learning classification reveals their physical causes.
Forecasting Volcanic Eruptions with Artificial Intelligence
A machine learning algorithm automatically detects telltale signs of volcanic unrest.
Machine Fault
Applying machine learning to subtle acoustic signals from an earthquake machine has revealed big clues about fault behavior in the lab.
Computers Tease Out Secrets of Jupiter’s Aurorae
Aurorae once classified by human eyes are now being sorted by machines. The change may help astronomers understand how the mysterious features are powered.
Virtual Super Instrument Enhances Solar Spacecraft
The same algorithms that help control self-driving cars and speech-to-text functionality have helped build a virtual instrument to study the Sun.