What am I looking at?
Two forecasts of the same weather, drawn on the same colour scale so they can be read against one another. MET Nordic is MET Norway's 1 km grid — the one behind Yr — statistically corrected against the observation network, published eight times a day to T+57. MEPS is the raw 2.5 km AROME run underneath it, shared with SMHI and FMI, out to T+66 and carrying the fields post-processing does not produce: fog, visibility, CAPE, precipitation type and the levels aloft.
Switching model holds the valid time, not the forecast hour, and every model here shares one time axis — hourly to T+36, three-hourly after. So flipping between the corrected 1 km grid and the raw 2.5 km run shows a real disagreement between two forecasts of the same instant, rather than an artefact of comparing two different moments on two different clocks.
Fields are drawn straight from MET Norway's published output with no smoothing and no rescaling — colour ramps are pinned to absolute values, so a loop shows real intensity changes rather than a palette re-fitting itself each frame. Temperatures are Celsius and winds metres per second, banded on the Norwegian wind-force scale, the way every forecast and every gale warning in the country states them. For warnings, yr.no and varsom.no are always the authority.