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

KMIA — Miami Intl

25.7906, -80.3164 · 3 m· standard offset UTC-5 h · truth product CLIMIA

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

the last 365 days, 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 4.05-4.05 · n=30 4.35-4.35 · n=29 4.40-4.40 · n=28 4.18-4.18 · n=27 4.32-4.32 · n=26 4.52-4.52 · n=25 4.44-4.44 · n=24 4.27-4.27 · n=23 4.21-4.21 · n=22 4.48-4.48 · n=21 ECMWF AIFS Single MAE by lead day 1 to 9, tmax
Aurora (GFS init)aurora_gfs 6.96-6.96 · n=2 6.93-6.93 · n=1 Aurora (GFS init) MAE by lead day 1 to 9, tmax
Aurora (IFS init)aurora_ifs 6.93-6.93 · n=3 6.88-6.88 · n=2 6.86-6.86 · n=1 Aurora (IFS init) MAE by lead day 1 to 9, tmax
FourCastNet v2 (GFS init)fourcastnet_gfs 4.54-4.54 · n=2 5.33-5.33 · n=1 FourCastNet v2 (GFS init) MAE by lead day 1 to 9, tmax
FourCastNet v2 (IFS init)fourcastnet_ifs 5.08-5.08 · n=3 4.98-4.98 · n=2 5.26-5.26 · n=1 FourCastNet v2 (IFS init) MAE by lead day 1 to 9, tmax
NCEP GFSgfs 1.52-1.25 · n=30 1.65-1.39 · n=29 1.69-1.46 · n=28 1.98-1.69 · n=27 2.04-1.91 · n=26 2.17-1.92 · n=25 2.71-2.54 · n=24 2.76-2.67 · n=23 2.56-2.56 · n=22 2.64-2.52 · n=21 NCEP GFS MAE by lead day 1 to 9, tmax
GraphCast (GFS init)graphcast_gfs 6.05-6.05 · n=2 6.08-6.08 · n=1 GraphCast (GFS init) MAE by lead day 1 to 9, tmax
GraphCast (IFS init)graphcast_ifs 3.48-3.48 · n=34 4.18-4.18 · n=33 4.40-4.40 · n=32 4.18-4.18 · n=31 4.31-4.31 · n=31 4.74-4.74 · n=31 4.76-4.73 · n=30 5.32-4.91 · n=29 5.18-5.14 · n=28 5.45-5.12 · n=27 GraphCast (IFS init) MAE by lead day 1 to 9, tmax
ECMWF IFS HRESifs_hres 3.27-3.27 · n=24 4.02-4.02 · n=23 3.95-3.95 · n=22 3.97-3.97 · n=22 3.57-3.57 · n=20 3.92-3.66 · n=18 3.95-3.85 · n=19 3.69-3.69 · n=18 4.27-4.15 · n=17 3.85-3.63 · n=16 ECMWF IFS HRES MAE by lead day 1 to 9, tmax
Pangu-Weather (GFS init)pangu_gfs 5.83-5.83 · n=2 5.99-5.99 · n=1 Pangu-Weather (GFS init) MAE by lead day 1 to 9, tmax
Pangu-Weather (IFS init)pangu_ifs 4.65-4.65 · n=3 6.55-6.55 · n=2 7.11-7.11 · n=1 Pangu-Weather (IFS init) MAE by lead day 1 to 9, tmax
Persistence (baseline)persistence 2.57+0.01 · n=365 3.38+0.01 · n=365 3.76+0.01 · n=365 3.97+0.02 · n=365 4.04+0.02 · n=365 4.02+0.02 · n=365 4.30+0.03 · n=365 4.36+0.02 · n=365 4.43+0.02 · n=365 Persistence (baseline) MAE by lead day 1 to 9, tmax

Up to 3 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

the last 365 days, 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 2.10+1.89 · n=30 1.89+1.60 · n=29 1.98+1.64 · n=28 1.92+1.63 · n=27 1.87+1.57 · n=26 2.01+1.61 · n=25 2.11+1.50 · n=24 2.13+1.65 · n=23 2.13+1.77 · n=22 2.40+1.98 · n=21 ECMWF AIFS Single MAE by lead day 1 to 9, tmin
Aurora (GFS init)aurora_gfs 2.72+1.99 · n=2 2.10-2.10 · n=1 Aurora (GFS init) MAE by lead day 1 to 9, tmin
Aurora (IFS init)aurora_ifs 1.84+1.49 · n=3 2.97+1.32 · n=2 1.72-1.72 · n=1 Aurora (IFS init) MAE by lead day 1 to 9, tmin
FourCastNet v2 (GFS init)fourcastnet_gfs 4.05+4.05 · n=2 0.61+0.61 · n=1 FourCastNet v2 (GFS init) MAE by lead day 1 to 9, tmin
FourCastNet v2 (IFS init)fourcastnet_ifs 2.31+2.26 · n=3 3.41+3.41 · n=2 0.73+0.73 · n=1 FourCastNet v2 (IFS init) MAE by lead day 1 to 9, tmin
NCEP GFSgfs 4.51+4.50 · n=30 4.84+4.78 · n=29 4.86+4.86 · n=28 4.92+4.92 · n=27 4.64+4.64 · n=26 4.80+4.80 · n=25 4.68+4.68 · n=24 4.76+4.76 · n=23 4.28+4.11 · n=22 4.60+4.60 · n=21 NCEP GFS MAE by lead day 1 to 9, tmin
GraphCast (GFS init)graphcast_gfs 3.60+3.60 · n=2 0.24-0.24 · n=1 GraphCast (GFS init) MAE by lead day 1 to 9, tmin
GraphCast (IFS init)graphcast_ifs 1.54+0.78 · n=34 1.71+0.27 · n=33 1.71+0.11 · n=32 1.42-0.28 · n=31 1.46-0.30 · n=31 1.82-0.49 · n=31 1.75-0.39 · n=30 2.13-1.15 · n=29 2.61-1.63 · n=28 3.86-1.66 · n=27 GraphCast (IFS init) MAE by lead day 1 to 9, tmin
ECMWF IFS HRESifs_hres 2.02+0.62 · n=24 1.68-0.17 · n=23 1.54-0.18 · n=22 1.65-0.04 · n=22 1.10+0.28 · n=20 1.25+0.61 · n=18 1.75+0.29 · n=19 1.86+0.53 · n=18 1.63+0.13 · n=17 1.89+0.68 · n=16 ECMWF IFS HRES MAE by lead day 1 to 9, tmin
Pangu-Weather (GFS init)pangu_gfs 3.53+3.53 · n=2 0.07-0.07 · n=1 Pangu-Weather (GFS init) MAE by lead day 1 to 9, tmin
Pangu-Weather (IFS init)pangu_ifs 1.32+1.27 · n=3 2.33+2.26 · n=2 0.74-0.74 · n=1 Pangu-Weather (IFS init) MAE by lead day 1 to 9, tmin
Persistence (baseline)persistence 2.74-0.01 · n=365 3.78-0.01 · n=365 4.26-0.01 · n=365 4.54-0.01 · n=365 4.84+0.00 · n=365 5.12-0.00 · n=365 5.36-0.00 · n=365 5.58-0.01 · n=365 5.63-0.02 · n=365 Persistence (baseline) 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, 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