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

KATL — Atlanta Hartsfield

33.6403, -84.4269 · 313 m· standard offset UTC-5 h · truth product CLIATL

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

all available history, 00Z, bilinear. MAE with the bias and n underneath; every cell links to its permanent page. The sparkline shows the same model across lead days 1–9 on a scale shared within this table.
Modeld0d1d2d3d4d5d6d7d8d9lead 1–9
ECMWF AIFS Singleaifs_single 4.36-4.36 · n=30 4.31-4.31 · n=29 4.38-4.37 · n=28 4.30-4.30 · n=27 4.21-4.21 · n=26 4.77-4.77 · n=25 4.89-4.68 · n=24 4.88-4.80 · n=23 5.53-5.53 · n=22 5.79-5.79 · n=21 ECMWF AIFS Single MAE by lead day 1 to 9, tmax
Aurora (GFS init)aurora_gfs 6.59-6.59 · n=2 7.11-7.11 · n=1 Aurora (GFS init) MAE by lead day 1 to 9, tmax
Aurora (IFS init)aurora_ifs 6.69-6.69 · n=3 6.26-6.26 · n=2 7.52-7.52 · n=1 Aurora (IFS init) MAE by lead day 1 to 9, tmax
FourCastNet v2 (GFS init)fourcastnet_gfs 3.67-3.67 · n=2 3.43-3.43 · n=1 FourCastNet v2 (GFS init) MAE by lead day 1 to 9, tmax
FourCastNet v2 (IFS init)fourcastnet_ifs 4.43-4.43 · n=3 2.78-2.78 · n=2 4.64-4.64 · n=1 FourCastNet v2 (IFS init) MAE by lead day 1 to 9, tmax
NCEP GFSgfs 1.17+0.15 · n=30 2.45-0.88 · n=29 3.13-0.78 · n=28 2.94+0.15 · n=27 3.26-0.72 · n=26 4.02-0.60 · n=25 4.91-0.67 · n=24 2.98-0.54 · n=23 3.71-0.27 · n=22 4.97-1.54 · n=21 NCEP GFS MAE by lead day 1 to 9, tmax
GraphCast (GFS init)graphcast_gfs 5.04-5.04 · n=2 6.18-6.18 · n=1 GraphCast (GFS init) MAE by lead day 1 to 9, tmax
GraphCast (IFS init)graphcast_ifs 5.67-5.65 · n=34 5.85-5.85 · n=33 5.86-5.86 · n=32 5.93-5.93 · n=31 5.52-5.51 · n=31 5.31-5.29 · n=31 5.99-5.78 · n=30 6.35-6.18 · n=29 6.29-5.66 · n=28 6.92-5.54 · n=27 GraphCast (IFS init) MAE by lead day 1 to 9, tmax
ECMWF IFS HRESifs_hres 2.64-2.38 · n=24 2.08-1.53 · n=23 2.57-1.55 · n=22 3.29-1.14 · n=22 4.00-2.08 · n=20 3.74-1.91 · n=18 3.12-0.12 · n=19 3.48-0.29 · n=18 2.62-0.28 · n=17 3.54-1.11 · n=16 ECMWF IFS HRES MAE by lead day 1 to 9, tmax
Pangu-Weather (GFS init)pangu_gfs 2.50-2.50 · n=2 3.98-3.98 · n=1 Pangu-Weather (GFS init) MAE by lead day 1 to 9, tmax
Pangu-Weather (IFS init)pangu_ifs 3.28-3.28 · n=3 2.86-2.86 · n=2 4.71-4.71 · n=1 Pangu-Weather (IFS init) MAE by lead day 1 to 9, tmax
Persistence (baseline)persistence 4.65-0.04 · n=972 6.27-0.08 · n=971 6.71-0.12 · n=970 7.11-0.16 · n=969 7.58-0.20 · n=968 7.78-0.25 · n=967 7.89-0.29 · n=966 8.12-0.33 · n=965 8.29-0.37 · n=964 Persistence (baseline) MAE by lead day 1 to 9, tmax

Up to 1 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

all available history, 00Z, bilinear. MAE with the bias and n underneath; every cell links to its permanent page. The sparkline shows the same model across lead days 1–9 on a scale shared within this table.
Modeld0d1d2d3d4d5d6d7d8d9lead 1–9
ECMWF AIFS Singleaifs_single 1.34-0.40 · n=30 1.48-0.27 · n=29 1.88-0.37 · n=28 1.70-0.46 · n=27 1.79-0.25 · n=26 1.94-0.14 · n=25 1.71-0.01 · n=24 1.93-0.04 · n=23 1.99-0.04 · n=22 2.16-0.08 · n=21 ECMWF AIFS Single MAE by lead day 1 to 9, tmin
Aurora (GFS init)aurora_gfs 1.27-1.27 · n=2 0.51-0.51 · n=1 Aurora (GFS init) MAE by lead day 1 to 9, tmin
Aurora (IFS init)aurora_ifs 1.87-1.87 · n=3 1.67-1.67 · n=2 1.31-1.31 · n=1 Aurora (IFS init) MAE by lead day 1 to 9, tmin
FourCastNet v2 (GFS init)fourcastnet_gfs 3.40-3.40 · n=2 0.09-0.09 · n=1 FourCastNet v2 (GFS init) MAE by lead day 1 to 9, tmin
FourCastNet v2 (IFS init)fourcastnet_ifs 3.33-3.33 · n=3 0.89-0.89 · n=2 0.21+0.21 · n=1 FourCastNet v2 (IFS init) MAE by lead day 1 to 9, tmin
NCEP GFSgfs 1.87+1.20 · n=30 2.15+1.42 · n=29 2.07+1.29 · n=28 2.40+1.70 · n=27 2.56+1.60 · n=26 2.97+2.01 · n=25 2.79+1.67 · n=24 2.70+1.08 · n=23 3.24+1.48 · n=22 3.85+0.81 · n=21 NCEP GFS MAE by lead day 1 to 9, tmin
GraphCast (GFS init)graphcast_gfs 1.43-1.43 · n=2 0.62-0.62 · n=1 GraphCast (GFS init) MAE by lead day 1 to 9, tmin
GraphCast (IFS init)graphcast_ifs 1.50-0.55 · n=34 1.60-0.48 · n=33 1.93-0.59 · n=32 2.03-0.82 · n=31 2.07-0.38 · n=31 2.24-0.06 · n=31 2.68-0.29 · n=30 3.08-0.49 · n=29 3.09-0.36 · n=28 3.11+0.94 · n=27 GraphCast (IFS init) MAE by lead day 1 to 9, tmin
ECMWF IFS HRESifs_hres 1.28-0.36 · n=24 1.93+0.19 · n=23 2.05+0.22 · n=22 2.55+0.62 · n=22 2.44+0.50 · n=20 2.24+0.47 · n=18 2.31+1.32 · n=19 2.74+1.18 · n=18 2.53+1.48 · n=17 2.66+1.72 · n=16 ECMWF IFS HRES MAE by lead day 1 to 9, tmin
Pangu-Weather (GFS init)pangu_gfs 1.35-1.35 · n=2 0.45-0.45 · n=1 Pangu-Weather (GFS init) MAE by lead day 1 to 9, tmin
Pangu-Weather (IFS init)pangu_ifs 1.45-1.45 · n=3 1.46-1.46 · n=2 0.83-0.83 · n=1 Pangu-Weather (IFS init) MAE by lead day 1 to 9, tmin
Persistence (baseline)persistence 3.94-0.04 · n=972 5.49-0.09 · n=971 6.36-0.13 · n=970 6.84-0.17 · n=969 7.15-0.21 · n=968 7.41-0.25 · n=967 7.66-0.28 · n=966 7.79-0.33 · n=965 7.88-0.36 · n=964 Persistence (baseline) MAE by lead day 1 to 9, tmin

Up to 1 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, model availability and the observed-truth table live on the canonical page: they do not depend on the window, initialization or interpolation selected here.

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