Otso 2.0

Otso 2.0 shows where fine particulate matter (PM2.5) will be over the next five days, at the scale of a square kilometre, on a street map.

Every morning, a deep-learning model runs on the day's atmospheric analysis of the Copernicus Atmosphere Monitoring Service, the ECMWF weather forecast and the fires seen by satellite, computes the transport of the air and forecasts PM2.5 from all its sources: traffic, industry, heating, agriculture, dust, sea salt and fires, over all land, as an ensemble of eight members. A second, generative model then draws the forecast at 1 km, trained on a satellite and station record of PM2.5.

• Search any city, street or address, or use your position, and read the forecast for that square kilometre.
• Five days ahead: the daily mean at 1 km, how the day evolves every 12 hours, and what it means in plain words.
• Three map layers: the 1 km daily mean, the number of ensemble members above the WHO 24-hour guideline (15 µg/m³), and the 40 km forecast every 12 hours.
• A table of sixty cities, to which you can add your own.
• Works offline: the last forecast of your place stays on your phone and is shown when there is no network.
• Dark mode, 37 languages, share a place with a link.
• Scores: how the forecast compares with air-quality monitors and with the Copernicus analysis, updated every day.

Otso is a research forecast from the Atmospheric Modelling Centre Lahti (University of Helsinki, LUT University). It is not an official air-quality service and does not replace the forecasts, indices and alerts of your national or local authorities. If you must decide something about your health, use those.

Contains modified Copernicus Atmosphere Monitoring Service information 2026; ECMWF open data (CC BY 4.0); street map © OpenMapTiles, data © OpenStreetMap contributors; city names © GeoNames (CC BY 4.0).