Where the weather models disagree in Minneapolis

The models are the independent global centres Open-Meteo serves for this point — GFS, ECMWF IFS, ICON, GEM, UKMO and the AIFS and AIGFS AI models — as The Spread recorded them at its last sweep. NBM and every other blend are left out: a consensus product hides the disagreement this page exists to show.

Each model publishes to a different horizon, which is why Models is counted per day rather than once. Beyond day 14 only two models remain and both are NOAA's, so the page stops there. Range is the lowest to the highest model high; Spread is the difference between those two printed numbers.

Typical spread is the median spread for that lead over the last month of forecasts made here, shown only once ten days of them exist. It is there because a wide band ten days out is ordinary and the same band tomorrow is not — so nothing on this page is called a divergence; the comparison is against what this lead usually looks like at this location.

GEFS p10–p90 is one centre's ensemble, shown beside the band and never as it: it verifies a few degrees warm at every lead.