TAF Trakerbeta
TAF performance
Every observation graded against the TAF in effect when it was taken. How grading works
Optimistic or cautious?
Each dot is an airport. Right = actual often worse than anything the TAF allowed. Up = actual often better than forecast.
Where TAFs miss
Share of observations worse than anything forecast. Larger dots = more data.
Wind above forecast, by lead time
Share of observations with wind or gusts 10+ kt above anything forecast. MOS forecasts sustained wind only.
Same category as forecast, by lead time
How often the forecast flight category was right, by how far ahead it was issued.
Worse than forecast, by lead time
Observed worse than anything forecast (including TEMPO/PROB).
Forecast vs. observed category
Station-hours, TAF in effect. Row = what the TAF said; column = what happened. Below the diagonal is worse than forecast.
Day by day
National share of observations worse than forecast, and same-category rate.
Time of day
Share of observations worse than forecast, by UTC hour.
All airports
Click a column to sort; click an airport for its detail page.
How grading works
Every 5-minute ASOS observation (or METAR/SPECI where an airport has no 5-minute feed) is compared with the TAF that was in effect at that moment: the prevailing conditions from its FM/BECMG chain, plus any TEMPO or PROB group active at that minute.
Flight category uses FAA limits on ceiling (lowest broken, overcast or vertical visibility) and visibility. Forecast ceilings above 12,000 ft are treated as "no ceiling" because ASOS can't see above that.