NOAA AIGFS Rainfall Forecast

About this data

• NOAA AIGFS forecast: NOAA AIGFS 0.25° deterministic (machine-learned forecast model) – U.S. Government work, public domain, via Open-Meteo Open Data
• Model runs: Four runs a day (00, 06, 12 and 18 GMT), each becoming available about five and a quarter hours after its nominal time, and every run reaches 16 days. AIGFS is NOAA's operational machine-learned model, built on Google DeepMind's GraphCast and started from the same analysis as GFS. It is published on a 0.25° (about 28km) grid in 6-hour steps: it smooths out showers and hill rain, the rate shown is the average over each six hours, the shortest total is 6 hours, and rainfall starts at the second step.
• Trial pipeline: this feed is built by a second, newer data pipeline (a small always-on server rather than the site's usual scheduled jobs), run alongside the existing one to see how well it holds up. The freshness dot above shows how old its newest data is.
• Deterministic, not a probability: this is a single model solution, so it shows one possible outcome rather than a range. Treat it as guidance on pattern and timing, not as a forecast of exact totals at a place.
• Comparing models: all seven render to the same map extent with the same colour scale, so a colour means the same millimetres whichever is selected. Differences you see between them are the models disagreeing, not the maps being drawn differently.
• Isobars (MSLP): off by default; switch on in the panel above (or in the bottom bar, on a wide enough screen) to add pressure lines over the rainfall. H/L pressure centres are labelled with their central pressure in hPa.
• Catchment boundaries: National River Flow Archive (2014). NERC Environmental Information Data Centre. doi:10.5285/1957166d.
• Region / county boundaries: Office for National Statistics (ONS), Open Government Licence v3.0. Contains OS data © Crown copyright and database right.
• Basemap: © OpenStreetMap contributors © CARTO.