How CAMS dust forecasts work, and how Nafas uses them

CAMS, the Copernicus Atmosphere Monitoring Service, is a European publicly-funded system that forecasts dust and other atmospheric pollutants worldwide, days ahead, by combining satellite observations with global atmospheric models. Nafas doesn't run its own forecasting model; it reads CAMS's dust and particulate output through the free Open-Meteo API, twice a day, for 10 Gulf cities, and turns the raw numbers into plain-language advice.

What CAMS is

CAMS is part of the Copernicus Programme, the European Union's Earth observation and environmental monitoring initiative. Its atmospheric-composition service produces regularly updated global forecasts covering pollutants, aerosols, and — relevant to the Gulf specifically — mineral dust, projecting conditions several days into the future. It's free and public, built for exactly the kind of downstream use Nafas makes of it: a data foundation other products build decision layers on top of, rather than a finished consumer forecast.

CAMS forecasts are model output, not a network of sensors reporting live readings. That distinction matters for how a user should read them: a forecast is a prediction, refreshed regularly as new observations feed back into the model, not a guarantee. It's also why cross-checking against satellite and ground readings, rather than treating the forecast as ground truth on its own, is built into how Nafas uses it.

How Nafas actually gets the data

Nafas doesn't call CAMS directly. It fetches CAMS's forecast through Open-Meteo's free air-quality API, which re-serves the same underlying Copernicus data in a simple, developer-friendly format. Twice a day, for each of the 10 Gulf cities it covers, Nafas requests an hourly PM10, PM2.5, and dust forecast covering the next three days, then keeps the next 48 hours of that window starting from the current hour. That's the entire "collect" step: no proprietary sensors, no paid data subscription, nothing installed in any of the 10 cities.

That forecast is cross-checked, not taken on faith. Nafas also reads Sentinel-5P satellite observations of aerosols and pollution gases and public ground-station readings via OpenAQ, so the CAMS forecast isn't the only signal in the system — it's the backbone Nafas builds the twice-daily update around.

The twice-daily cadence is a deliberate choice, not a limit of what's available. CAMS's underlying data refreshes more often than that behind the scenes, but Nafas re-fetches and rebuilds its city-by-city forecast twice a day, which keeps every advice line grounded in a recent run without requiring a user to keep checking back constantly. A stale-data banner on the live demo makes it obvious if a refresh has been missed, rather than silently serving an outdated forecast.

What Nafas adds on top

The forecast itself is CAMS's work, not Nafas's. What Nafas adds is the layer after it: scoring every forecast hour against US EPA AQI breakpoints, finding the worst stretch of hours in the coming day, and writing that up as one plain sentence instead of a table of numbers — the same rule-based process covered in the forecast-to-decision explainer. That's a deliberately modest, well-understood layer today. A Gulf-specific dust forecasting model, tuned on Nafas's own daily forecast/observation validation archive rather than trained on global defaults, is on the roadmap, not something already running in production — see the About page for where that stands.

In short: CAMS tells Nafas what the atmosphere over a city is likely to do next. Nafas's job is turning that into something a facilities manager or school administrator can act on without reading a forecast chart. Checktoday's live forecast to see the current output for any of the 10 cities.