Get an ECMWF weather forecast with AI
Look up the forecast for a location and distinguish the model’s run time from the time it predicts. Get values from the nearest available grid point.
To get an ECMWF weather forecast with AI, give a location, a time window and the quantity you need. A team planning an outdoor recording might ask for the wind at the start of the session and total rain during the hours they will be outside.
The ecmwf-open-data skill finds a suitable model run and reads the relevant grid point. It can distinguish accumulated rainfall from the amount in an interval and identify the time, model and source behind the result.
What it does
- Ask for the temperature at a site over the hours of an outdoor event.
- Compare wind speed at two locations using the same forecast time.
- Calculate the rain expected during a chosen interval instead of quoting the cumulative total.
- Record the model run and nearest grid point alongside a planning note.
- Estimate the download size before retrieving selected fields from a forecast file.
Install and use
- Install the skill and make Python available. Point decoding uses uv and earthkit packages. The public Open Data route does not require an account.
- Ask, “Find the wind and rain for this location between eight and noon tomorrow in local time.”
- Check the run, validity period and units, and preserve the data credit when using the figures elsewhere.
FAQ
Source
Made by ECMWF. Original skill: ecmwf-open-data, published on GitHub under the Apache-2.0 licence.
View on GitHub