Raw model output on the native 0.25° grid — no MOS, no bias correction, no post-processing. Scores understate operational forecast quality. Read the fairness statement →

CastCheck methodology v0.2 · data through 2026-08-30

GraphCast (GFS init) (graphcast_gfs)

source aiwp · product GRAP · initial conditions from GFS · steps every 6 h to 240 h · initializations 0, 12 UTC.

Scores use the latest model version segment, GRAP_v100, starting 2026-08-28. Earlier days ran a different version of the model and are excluded rather than silently pooled into the same average.

These pages are always bilinear interpolation, the headline method. The nearest-node variant, and every other window/initialization/interpolation/variable combination, is published on each permanent link below — follow any number.

Daily maximum temperature — MAE °F by station

the last 30 days, 00Z, bilinear. MAE with the bias and n underneath; every cell links to its permanent page. The sparkline shows this station across lead days 1–9 on a scale shared within this table.
Stationd0d1d2d3d4d5d6d7d8d9lead 1–9
KATL 5.04-5.04 · n=2 6.18-6.18 · n=1 KATL MAE by lead day 1 to 9, tmax
KAUS 6.93-6.93 · n=2 8.63-8.63 · n=1 KAUS MAE by lead day 1 to 9, tmax
KBOS 4.38-4.38 · n=2 5.34-5.34 · n=1 KBOS MAE by lead day 1 to 9, tmax
KDCA 5.94-5.94 · n=2 5.13-5.13 · n=1 KDCA MAE by lead day 1 to 9, tmax
KDEN 5.27-5.27 · n=2 3.06-3.06 · n=1 KDEN MAE by lead day 1 to 9, tmax
KDFW 4.86-4.86 · n=2 5.83-5.83 · n=1 KDFW MAE by lead day 1 to 9, tmax
KEWR 3.88-3.88 · n=2 4.59-4.59 · n=1 KEWR MAE by lead day 1 to 9, tmax
KIAH 4.09-4.09 · n=2 5.55-5.55 · n=1 KIAH MAE by lead day 1 to 9, tmax
KLAS 7.94-7.94 · n=2 7.64-7.64 · n=1 KLAS MAE by lead day 1 to 9, tmax
KLAX 5.57-5.57 · n=2 5.25-5.25 · n=1 KLAX MAE by lead day 1 to 9, tmax
KMIA 6.05-6.05 · n=2 6.08-6.08 · n=1 KMIA MAE by lead day 1 to 9, tmax
KMSP 6.98-6.98 · n=2 7.60-7.60 · n=1 KMSP MAE by lead day 1 to 9, tmax
KMSY 5.57-5.57 · n=2 5.32-5.32 · n=1 KMSY MAE by lead day 1 to 9, tmax
KNYC 0.97-0.97 · n=2 2.18-2.18 · n=1 KNYC MAE by lead day 1 to 9, tmax
KOKC 5.74-5.74 · n=2 7.68-7.68 · n=1 KOKC MAE by lead day 1 to 9, tmax
KORD 3.52-3.52 · n=2 6.94-6.94 · n=1 KORD MAE by lead day 1 to 9, tmax
KPHL 5.26-5.26 · n=2 6.35-6.35 · n=1 KPHL MAE by lead day 1 to 9, tmax
KPHX 14.89-14.89 · n=2 13.90-13.90 · n=1 KPHX MAE by lead day 1 to 9, tmax
KSAN 5.30-5.30 · n=2 3.50-3.50 · n=1 KSAN MAE by lead day 1 to 9, tmax
KSAT 6.67-6.67 · n=2 7.52-7.52 · n=1 KSAT MAE by lead day 1 to 9, tmax
KSEA 6.13-6.13 · n=2 4.43-4.43 · n=1 KSEA MAE by lead day 1 to 9, tmax
KSFO 6.19-6.19 · n=2 4.64-4.64 · n=1 KSFO MAE by lead day 1 to 9, tmax
KTTN 1.76-1.76 · n=2 3.44-3.44 · n=1 KTTN MAE by lead day 1 to 9, tmax

Up to 2 of the scored days in this table carry a QC flag on the observation (a CLI value that disagreed with the hourly observations, or a fallback source). They are kept in the scores and counted here rather than dropped — see methodology.

+ model too warm · − model too cold · bias interval includes zero · ★ lowest MAE · = not distinguishable from the leader · ▼ significantly worse · n < 30 greyed and unranked · CI is a 95 % moving-block bootstrap interval.

Daily minimum temperature — MAE °F by station

the last 30 days, 00Z, bilinear. MAE with the bias and n underneath; every cell links to its permanent page. The sparkline shows this station across lead days 1–9 on a scale shared within this table.
Stationd0d1d2d3d4d5d6d7d8d9lead 1–9
KATL 1.43-1.43 · n=2 0.62-0.62 · n=1 KATL MAE by lead day 1 to 9, tmin
KAUS 6.49+6.49 · n=2 4.84+4.84 · n=1 KAUS MAE by lead day 1 to 9, tmin
KBOS 3.50-3.50 · n=2 4.45-4.45 · n=1 KBOS MAE by lead day 1 to 9, tmin
KDCA 2.09-2.09 · n=2 1.22-1.22 · n=1 KDCA MAE by lead day 1 to 9, tmin
KDEN 4.74+4.74 · n=2 3.27+3.27 · n=1 KDEN MAE by lead day 1 to 9, tmin
KDFW 2.50+2.50 · n=2 0.85+0.85 · n=1 KDFW MAE by lead day 1 to 9, tmin
KEWR 1.83+1.83 · n=2 2.38+2.38 · n=1 KEWR MAE by lead day 1 to 9, tmin
KIAH 2.29+2.29 · n=2 0.93+0.93 · n=1 KIAH MAE by lead day 1 to 9, tmin
KLAS 6.58-6.58 · n=2 9.42-9.42 · n=1 KLAS MAE by lead day 1 to 9, tmin
KLAX 1.21-1.21 · n=2 2.09-2.09 · n=1 KLAX MAE by lead day 1 to 9, tmin
KMIA 3.60+3.60 · n=2 0.24-0.24 · n=1 KMIA MAE by lead day 1 to 9, tmin
KMSP 1.24+1.24 · n=2 1.36-1.36 · n=1 KMSP MAE by lead day 1 to 9, tmin
KMSY 2.51-0.22 · n=2 2.65-2.65 · n=1 KMSY MAE by lead day 1 to 9, tmin
KNYC 0.59+0.45 · n=2 0.26+0.26 · n=1 KNYC MAE by lead day 1 to 9, tmin
KOKC 7.30+7.30 · n=2 5.73+5.73 · n=1 KOKC MAE by lead day 1 to 9, tmin
KORD 0.75+0.75 · n=2 0.90+0.90 · n=1 KORD MAE by lead day 1 to 9, tmin
KPHL 0.45-0.45 · n=2 1.56+1.56 · n=1 KPHL MAE by lead day 1 to 9, tmin
KPHX 6.39-6.39 · n=2 8.09-8.09 · n=1 KPHX MAE by lead day 1 to 9, tmin
KSAN 1.52-0.94 · n=2 1.97-1.97 · n=1 KSAN MAE by lead day 1 to 9, tmin
KSAT 1.07-0.46 · n=2 2.20-2.20 · n=1 KSAT MAE by lead day 1 to 9, tmin
KSEA 1.21+1.21 · n=2 0.46+0.46 · n=1 KSEA MAE by lead day 1 to 9, tmin
KSFO 0.62+0.12 · n=2 0.89-0.89 · n=1 KSFO MAE by lead day 1 to 9, tmin
KTTN 3.99+3.99 · n=2 6.61+6.61 · n=1 KTTN MAE by lead day 1 to 9, tmin

Up to 2 of the scored days in this table carry a QC flag on the observation (a CLI value that disagreed with the hourly observations, or a fallback source). They are kept in the scores and counted here rather than dropped — see methodology.

Month-by-month slices live on the canonical page: they do not depend on the window, initialization or interpolation selected here.

Source and licence

Where it comes from

aiwp · GRAP · GFS initial conditions

Steps every 6 h to 240 h; sampled at 00/06/12/18 UTC only.

Licence

NOAA/CIRA AIWP, U.S. Government work, public domain. Cite Radford et al. (2025), BAMS, doi:10.1175/BAMS-D-24-0057.1.

Post-processing

None. Raw grid values, bilinear (headline) or nearest-node interpolation, no MOS, no bias correction, no downscaling.

All scores as JSON scores_latest.csv daily_errors.csv.gz data & licences