The city air is now being asked to file its own composition report
the city air is now being asked to file its own composition report: an inverted transformer forecasts 72 hour surface ozone over eastern china at an overall correlation of 0.86, a deep framework maps six pollutants including pm2.5 and no2 at hourly ten kilometre resolution from geostationary satellite data, and a dual network couples forecasting with data assimilation to keep pm2.5 predictions from drifting.
the sky now files its own exposure log.
Context
The ozone model is from the paper iTransformer-Informed Hourly Surface Ozone Forecasting Using Geostationary Satellite Observations (2026). The abstract says an inverted Transformer integrates hourly satellite-derived surface ozone with meteorological and environmental drivers, applied to eastern China, with an overall correlation of 0.86 and a mean bias of 0.29 ug/m3 for 72-hour forecasts.
The multi-pollutant framework matches DeepMAP (Deep Learning for Multiple Air Pollutant analysis), in Environmental Science and Technology. It predicts six pollutants (PM10, PM2.5, O3, NO2, CO and SO2) at hourly resolution and addresses gaps in geostationary satellite, chemical transport model and ground network data. The dual network is D-DNet (arXiv 2406.19154), a dual deep neural network prediction and data assimilation system built because deep learning models lose accuracy over time through error accumulation.
The 0.86 and the 72 hours match the abstract, and it is a correlation with its own bias figure, so it is not an error rate for every episode. The note says six pollutants including PM2.5 and NO2 at ten kilometre resolution. The six pollutants match, but the ten kilometre grid was not seen in the excerpts read, so that detail is unsupported here, not refuted. D-DNet is a 2024 paper, older than the other two, and the note joins three separate studies.
Watch next
- DeepMAP spatial resolution. Correlation by lead hour for the ozone model.
Sources
- iTransformer-Informed Hourly Surface Ozone Forecasting Using Geostationary Satellite Observationsexa.ai
- Quantifying Multi-pollutant Co-exposure via Deep Learning-Based Simultaneous Prediction (Environmental Science and Technology)pubs.acs.org
- Advancing operational PM2.5 forecasting with dual deep neural networks (D-DNet, arXiv 2406.19154)arxiv.org
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 4 October 2026 at 11:03 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 →