The volcano is now being asked to file its own heat report
the volcano is now being asked to file its own heat report: a random forest on sentinel 2 spots thermal anomalies from orbit without fixed thresholds, a neural net tells ash ice and sulfur dioxide clouds apart on seviri images of etna, and a cnn separates no eruption, phreatic and magmatic events at 91.2 percent accuracy.
the mountain now files its own unrest notice.
Context
Data-Driven Random Forest Models for Detecting Volcanic Thermal Anomalies (September 2, 2022) uses Sentinel-2 spectral bands and indices to map lava-related thermal anomalies. Its reported area-overlap accuracy is around 0.9, compared with 0.83 for the fixed-threshold method, across the study cases. Targets were drawn through expert visual inspection.
Characterization of Volcanic Cloud Components Using Machine Learning Techniques and SEVIRI Infrared Images (2022) uses a supervised support vector machine in Google Earth Engine. It classifies Etna cloud pixels as ash, sulfur dioxide, mixed ash and sulfur dioxide, or background. Testing uses three images not used for training; cloud-component Micro-F1 is 0.86 and Macro-F1 is 0.83.
The located SEVIRI study uses a support vector machine, not a neural network, and does not define ice as a separate output class. This does not refute a different neural-network study, but that exact claim remains unsupported by the sources read. The random-forest accuracy is an area intersection-over-union measure, not overall pixel accuracy.
A separate 2021 on-board CNN feasibility study reports binary eruption versus non-eruption test accuracy of 85% for the larger model and 83% for the smaller model. It does not establish the note’s three-class, 91.2% claim, which remains unsupported here, not refuted. Thermal mapping, cloud composition and eruption classification are distinct tasks.
Watch next
- The primary paper behind the three-class 91.2% result and the proposed neural-network ice classifier. Tests across unseen volcanoes, cloud conditions and prospective unrest episodes.
Sources
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 4 October 2026 at 07:43 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →