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

Miami Intl · NCEP GFS · lead day 3

KMIA · gfs · lead 3

Target day = initialization date + 3 days. This page is a permanent link: the numbers change as history accumulates, the address and the methodology version do not.

Scores use the latest model version segment, gfs-0p25, starting 2026-08-01 — days before that ran a different version of the model and are excluded rather than silently pooled. Of the scored days, 2 carry a QC flag on the observation.

JSON for this page CSV of the daily errors Cite this page

Every number for this combination

All four windows × both initializations × both interpolations × both variables. Errors are computed in °C and shown in °F; CI is a 95 % moving-block bootstrap interval. MAE debiased and skill, debiased remove the constant offset of the window first, so a station that the model always reads a degree too warm is not scored as if it had no skill — compare the two skill columns to see how much of the error is a fixed offset. n QC next to n counts days whose observation carries a QC flag.

VariableInitMethod Windown MAE °F MAE debiased Bias °F RMSE °F ±1 °F±2 °F±3 °F Skill Skill, debiased Period
tmax00Zbilinear30d27 2 QC1.98 [1.46, 2.45]0.98-1.69 [-2.40, -0.79]2.20 [1.55, 2.68]7% [0%, 22%]63%89%+0.08+0.542026-08-04 → 2026-08-30
tmax00Zbilinear90d27 2 QC1.98 [1.33, 2.44]0.98-1.69 [-2.40, -0.49]2.20 [1.44, 2.67]7% [0%, 25%]63%89%+0.08+0.542026-08-04 → 2026-08-30
tmax00Zbilinear365d27 2 QC1.98 [1.33, 2.49]0.98-1.69 [-2.45, -0.51]2.20 [1.42, 2.71]7% [0%, 25%]63%89%+0.08+0.542026-08-04 → 2026-08-30
tmax00Zbilinearall27 2 QC1.98 [1.29, 2.46]0.98-1.69 [-2.43, -0.45]2.20 [1.38, 2.69]7% [0%, 28%]63%89%+0.08+0.542026-08-04 → 2026-08-30
tmax00Znearest30d27 2 QC3.53 [2.38, 4.45]1.23-3.53 [-4.45, -2.38]3.86 [2.76, 4.60]7% [0%, 17%]15%26%-0.64+0.432026-08-04 → 2026-08-30
tmax00Znearest90d27 2 QC3.53 [2.02, 4.44]1.23-3.53 [-4.44, -2.02]3.86 [2.39, 4.60]7% [0%, 21%]15%26%-0.64+0.432026-08-04 → 2026-08-30
tmax00Znearest365d27 2 QC3.53 [2.02, 4.53]1.23-3.53 [-4.53, -2.02]3.86 [2.39, 4.68]7% [0%, 20%]15%26%-0.64+0.432026-08-04 → 2026-08-30
tmax00Znearestall27 2 QC3.53 [1.90, 4.49]1.23-3.53 [-4.49, -1.90]3.86 [2.24, 4.63]7% [0%, 22%]15%26%-0.64+0.432026-08-04 → 2026-08-30
tmax12Zbilinear30d27 2 QC2.09 [1.42, 2.61]1.18-1.88 [-2.60, -0.91]2.43 [1.73, 2.89]19% [4%, 39%]56%78%+0.03+0.452026-08-04 → 2026-08-30
tmax12Zbilinear90d27 2 QC2.09 [1.24, 2.56]1.18-1.88 [-2.56, -0.66]2.43 [1.50, 2.87]19% [6%, 40%]56%78%+0.03+0.452026-08-04 → 2026-08-30
tmax12Zbilinear365d27 2 QC2.09 [1.26, 2.62]1.18-1.88 [-2.62, -0.68]2.43 [1.50, 2.92]19% [5%, 42%]56%78%+0.03+0.452026-08-04 → 2026-08-30
tmax12Zbilinearall27 2 QC2.09 [1.12, 2.61]1.18-1.88 [-2.61, -0.46]2.43 [1.30, 2.89]19% [5%, 47%]56%78%+0.03+0.452026-08-04 → 2026-08-30
tmax12Znearest30d27 2 QC3.79 [2.66, 4.65]1.32-3.79 [-4.65, -2.66]4.13 [3.08, 4.80]7% [0%, 17%]15%30%-0.76+0.382026-08-04 → 2026-08-30
tmax12Znearest90d27 2 QC3.79 [2.41, 4.63]1.32-3.79 [-4.63, -2.41]4.13 [2.79, 4.79]7% [0%, 21%]15%30%-0.76+0.382026-08-04 → 2026-08-30
tmax12Znearest365d27 2 QC3.79 [2.45, 4.69]1.32-3.79 [-4.69, -2.45]4.13 [2.81, 4.86]7% [0%, 20%]15%30%-0.76+0.382026-08-04 → 2026-08-30
tmax12Znearestall27 2 QC3.79 [2.14, 4.69]1.32-3.79 [-4.69, -2.14]4.13 [2.50, 4.82]7% [0%, 22%]15%30%-0.76+0.382026-08-04 → 2026-08-30
tmin00Zbilinear30d27 2 QC4.92 [3.87, 6.10]1.87+4.92 [3.87, 6.10]5.47 [4.52, 6.44]4% [0%, 10%]7%26%-1.01+0.242026-08-04 → 2026-08-30
tmin00Zbilinear90d27 2 QC4.92 [3.77, 6.31]1.87+4.92 [3.77, 6.31]5.47 [4.40, 6.59]4% [0%, 11%]7%26%-1.01+0.242026-08-04 → 2026-08-30
tmin00Zbilinear365d27 2 QC4.92 [3.74, 6.55]1.87+4.92 [3.74, 6.55]5.47 [4.39, 6.76]4% [0%, 11%]7%26%-1.01+0.242026-08-04 → 2026-08-30
tmin00Zbilinearall27 2 QC4.92 [3.73, 6.34]1.87+4.92 [3.73, 6.34]5.47 [4.36, 6.63]4% [0%, 11%]7%26%-1.01+0.242026-08-04 → 2026-08-30
tmin00Znearest30d27 2 QC5.24 [4.11, 6.45]1.89+5.24 [4.11, 6.45]5.77 [4.74, 6.77]4% [0%, 10%]7%19%-1.14+0.232026-08-04 → 2026-08-30
tmin00Znearest90d27 2 QC5.24 [4.00, 6.64]1.89+5.24 [4.00, 6.64]5.77 [4.64, 6.90]4% [0%, 11%]7%19%-1.14+0.232026-08-04 → 2026-08-30
tmin00Znearest365d27 2 QC5.24 [3.97, 6.86]1.89+5.24 [3.97, 6.86]5.77 [4.61, 7.10]4% [0%, 11%]7%19%-1.14+0.232026-08-04 → 2026-08-30
tmin00Znearestall27 2 QC5.24 [3.95, 6.69]1.89+5.24 [3.95, 6.69]5.77 [4.60, 6.95]4% [0%, 11%]7%19%-1.14+0.232026-08-04 → 2026-08-30
tmin12Zbilinear30d27 2 QC4.49 [3.55, 5.65]1.92+4.49 [3.55, 5.65]5.10 [4.12, 6.08]4% [0%, 10%]11%33%-0.84+0.222026-08-04 → 2026-08-30
tmin12Zbilinear90d27 2 QC4.49 [3.42, 5.84]1.92+4.49 [3.42, 5.84]5.10 [4.07, 6.19]4% [0%, 11%]11%33%-0.84+0.222026-08-04 → 2026-08-30
tmin12Zbilinear365d27 2 QC4.49 [3.41, 6.09]1.92+4.49 [3.41, 6.09]5.10 [3.97, 6.44]4% [0%, 11%]11%33%-0.84+0.222026-08-04 → 2026-08-30
tmin12Zbilinearall27 2 QC4.49 [3.35, 5.84]1.92+4.49 [3.35, 5.84]5.10 [3.92, 6.21]4% [0%, 11%]11%33%-0.84+0.222026-08-04 → 2026-08-30
tmin12Znearest30d27 2 QC4.81 [3.81, 5.96]1.92+4.81 [3.81, 5.96]5.38 [4.41, 6.35]4% [0%, 10%]7%30%-0.97+0.222026-08-04 → 2026-08-30
tmin12Znearest90d27 2 QC4.81 [3.70, 6.13]1.92+4.81 [3.70, 6.13]5.38 [4.30, 6.45]4% [0%, 11%]7%30%-0.97+0.222026-08-04 → 2026-08-30
tmin12Znearest365d27 2 QC4.81 [3.69, 6.38]1.92+4.81 [3.69, 6.38]5.38 [4.25, 6.67]4% [0%, 11%]7%30%-0.97+0.222026-08-04 → 2026-08-30
tmin12Znearestall27 2 QC4.81 [3.64, 6.14]1.92+4.81 [3.64, 6.14]5.38 [4.15, 6.51]4% [0%, 11%]7%30%-0.97+0.222026-08-04 → 2026-08-30

+ 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 error, last 90 days

Daily forecast minus observed error in °F for NCEP GFS at KMIA, lead day 3, 00Z bilinear tmax+5+2.50-2.5-52026-08-04: -1.66 °F2026-08-05: +0.38 °F2026-08-06: +1.39 °F2026-08-07: -1.89 °F2026-08-08: -1.78 °F2026-08-09: -2.19 °F2026-08-10: -2.79 °F2026-08-11: -3.48 °F2026-08-12: -2.11 °F2026-08-13: -1.54 °F2026-08-14: -4.32 °F2026-08-15: -2.43 °F2026-08-16: -2.58 °F2026-08-17: -1.58 °F2026-08-18: -4.72 °F2026-08-19: -1.59 °F2026-08-20: -1.1 °F2026-08-21: -1.97 °F2026-08-22: -2.22 °F2026-08-23: -2.48 °F2026-08-24: -1.15 °F2026-08-25: -1.1 °F2026-08-26: -1.72 °F2026-08-27: -1.22 °F2026-08-28: -1.83 °F2026-08-29: +1.52 °F2026-08-30: +0.61 °F2026-08-042026-08-30°F, forecast − observed
Forecast − observed, °F, 00Z bilinear daily maximum. The dashed line is zero; the thick line is a centred 7-day mean, drawn only where the whole window is present. Gaps are days with no score, never interpolation. Hover a point for its date and value.
The same series as a table (27 days, newest first) · download CSV
Climatological dayError °F
2026-08-30+0.61
2026-08-29+1.52
2026-08-28-1.83
2026-08-27-1.22
2026-08-26-1.72
2026-08-25-1.10
2026-08-24-1.15
2026-08-23-2.48
2026-08-22-2.22
2026-08-21-1.97
2026-08-20-1.10
2026-08-19-1.59
2026-08-18-4.72
2026-08-17-1.58
2026-08-16-2.58
2026-08-15-2.43
2026-08-14-4.32
2026-08-13-1.54
2026-08-12-2.11
2026-08-11-3.48
2026-08-10-2.79
2026-08-09-2.19
2026-08-08-1.78
2026-08-07-1.89
2026-08-06+1.39
2026-08-05+0.38
2026-08-04-1.66

Error distribution

Distribution of the signed daily error in °F for NCEP GFS at KMIA, lead day 3-5 to -4.5 °F: 1 day-4.5 to -4 °F: 1 day-4 to -3.5 °F: 0 days-3.5 to -3 °F: 1 day-3 to -2.5 °F: 2 days-2.5 to -2 °F: 5 days-2 to -1.5 °F: 9 days-1.5 to -1 °F: 4 days-1 to -0.5 °F: 0 days-0.5 to 0 °F: 0 days0 to 0.5 °F: 1 day0.5 to 1 °F: 1 day1 to 1.5 °F: 1 day1.5 to 2 °F: 1 day2 to 2.5 °F: 0 days2.5 to 3 °F: 0 days3 to 3.5 °F: 0 days3.5 to 4 °F: 0 days4 to 4.5 °F: 0 days4.5 to 5 °F: 0 daysdashed: P50 |e| 1.78 · P90 |e| 3.06 °F-50+5 °Fn = 27 days
All scored days for 00Z bilinear daily maximum, 20 equal bins. The dashed marks are the 50th and 90th percentiles of the absolute error, mirrored on both sides. Hover a bar for its bin and count.
Bin counts as a table
From °FTo °FDays
-5.00-4.501
-4.50-4.001
-4.00-3.500
-3.50-3.001
-3.00-2.502
-2.50-2.005
-2.00-1.509
-1.50-1.004
-1.00-0.500
-0.50+0.000
+0.00+0.501
+0.50+1.001
+1.00+1.501
+1.50+2.001
+2.00+2.500
+2.50+3.000
+3.00+3.500
+3.50+4.000
+4.00+4.500
+4.50+5.000

Month by month

00Z bilinear daily maximum, newest first.

Monthn MAE °FBias °F
2026-08271.98 -1.69

Availability

ModelScored period DaysCoverage
NCEP GFS2026-08-04 → 2026-08-30 27
NCEP GFS2026-08-04 → 2026-08-30 27
NCEP GFS2026-08-04 → 2026-08-30 27
NCEP GFS2026-08-04 → 2026-08-30 27

Compared with other models

Paired bootstrap on the days both models cover (90d, 00Z, bilinear, daily maximum). Negative means NCEP GFS has the lower MAE. "Distinguishable" means the 95 % interval excludes zero.

Other modelCommon days MAE difference °FDistinguishable
ECMWF AIFS Single 27 -2.21 [-3.44, -1.47] yes
GraphCast (IFS init) 25 -2.41 [-4.44, -1.16] yes
ECMWF IFS HRES 22 -1.89 [-3.24, -0.91] yes
Persistence (baseline) 27 -0.17 [-0.79, 0.52] no

Cite this page

ShortCastCheck, KMIA gfs lead 3, methodology v0.2, accessed 2026-08-31, https://castcheck.zifanzhang.com/station/KMIA/model/gfs/lead/3/

LongCastCheck (2026). Station-level verification of raw weather-model 2 m temperature forecasts: KMIA (Miami Intl), model gfs, lead day 3. Methodology version 0.2, schema version 0.1, data through 2026-08-30. Accessed 2026-08-31. https://castcheck.zifanzhang.com/station/KMIA/model/gfs/lead/3/

Accessed2026-08-31 (data through 2026-08-30; page built 2026-08-31T03:52:06+00:00)

JSON CSV all scores methodology v0.2 licence