EXIT 50 · Downscaled · Weather Trader / ECMWF
AIFS-DS map explorer
The AIFS downscaled map feed on the 2.5 km URMA grid, six-hourly to 360 hours. Its gridded products and history are available in the explorer; no full-day station-market score is inferred from map colors.
What to inspect
Follow successive runs for the same target date. Compare error, bias and interval coverage on shared stations and dates. A sharper forecast is useful only when its uncertainty remains reliable.
Separate native-window guidance; no compatible full-day market score
Updates and availability
See issued guidance for available cycles; collection and upstream release times differ
No issue timestamp is exposed by this directory for this product. The linked guidance page carries its own coverage and run information.
Station and neighborhood context
Use the station page to compare diurnal evolution, nearby observations and weather regimes. The linked page contains all available sources; availability varies by station.
How to read the record
Raw models, corrected station products and downscaled grids forecast different things. Market recap shows its decision cutoff, settlement basis and coverage. Missing forecasts are unavailable, never zero. Inclusion in this directory or recap does not imply inclusion in a trading blend.
Existing model verification and cycle archive ↗Related exits
EXIT 03 ↗GlobalECMWF AIFS
A machine-learned global forecast. Keep its native forecast separate from the station-corrected AIFS-MOS product.
ECMWFExplore guidance and its verification recordEXIT 04 ↗EnsembleAIFS-ENS
The ensemble counterpart to the AI forecast. Its pattern verification and station-market coverage are separate records.
ECMWFExplore guidance and its verification recordEXIT 24 ↗MOSAIFS-MOS
Daily station highs and lows derived from ECMWF AIFS Single, corrected using walk-forward station equations. This is the product sometimes called AIF-MOS.
Weather Trader / ECMWFIssue record: 2026-10-18T00:15:21ZEXIT 49 ↗DownscaledAIFS-DS 2.5 km
ECMWF AIFS Single guidance corrected against URMA on a 2.5 km grid with a station layer, the same method as ECMWF-DS. Daily highs and lows are rebuilt from its 6-hourly temperature. Native-window verification is distinct from full-day market verification.
Weather Trader / ECMWFIssue record: 2026-10-17T23:42Z