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

Atlanta Hartsfield · NCEP GFS · lead day 0

KATL · gfs · lead 0

Target day = initialization date + 0 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, 1 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
tmax00Zbilinear30d30 1 QC1.17 [0.95, 1.39]1.16+0.15 [-0.31, 0.63]1.49 [1.28, 1.67]53% [37%, 70%]77%97%——2026-08-01 → 2026-08-30
tmax00Zbilinear90d30 1 QC1.17 [0.95, 1.44]1.16+0.15 [-0.43, 0.59]1.49 [1.29, 1.67]53% [31%, 71%]77%97%——2026-08-01 → 2026-08-30
tmax00Zbilinear365d30 1 QC1.17 [0.91, 1.42]1.16+0.15 [-0.41, 0.62]1.49 [1.26, 1.67]53% [31%, 72%]77%97%——2026-08-01 → 2026-08-30
tmax00Zbilinearall30 1 QC1.17 [0.92, 1.45]1.16+0.15 [-0.45, 0.61]1.49 [1.24, 1.68]53% [29%, 71%]77%97%——2026-08-01 → 2026-08-30
tmax00Znearest30d30 1 QC1.38 [1.15, 1.58]1.30+0.62 [0.22, 1.00]1.69 [1.46, 1.86]43% [30%, 57%]77%90%——2026-08-01 → 2026-08-30
tmax00Znearest90d30 1 QC1.38 [1.16, 1.58]1.30+0.62 [0.12, 0.99]1.69 [1.44, 1.87]43% [29%, 58%]77%90%——2026-08-01 → 2026-08-30
tmax00Znearest365d30 1 QC1.38 [1.13, 1.60]1.30+0.62 [0.11, 1.02]1.69 [1.42, 1.89]43% [29%, 59%]77%90%——2026-08-01 → 2026-08-30
tmax00Znearestall30 1 QC1.38 [1.13, 1.59]1.30+0.62 [0.11, 1.02]1.69 [1.41, 1.88]43% [29%, 59%]77%90%——2026-08-01 → 2026-08-30
tmin00Zbilinear30d30 1 QC1.87 [1.42, 2.29]1.59+1.20 [0.17, 2.14]2.42 [1.72, 2.99]33% [23%, 47%]67%80%——2026-08-01 → 2026-08-30
tmin00Zbilinear90d30 1 QC1.87 [1.34, 2.27]1.59+1.20 [-0.15, 2.06]2.42 [1.61, 3.00]33% [23%, 48%]67%80%——2026-08-01 → 2026-08-30
tmin00Zbilinear365d30 1 QC1.87 [1.39, 2.33]1.59+1.20 [-0.18, 2.17]2.42 [1.64, 3.03]33% [21%, 50%]67%80%——2026-08-01 → 2026-08-30
tmin00Zbilinearall30 1 QC1.87 [1.30, 2.26]1.59+1.20 [-0.12, 2.12]2.42 [1.55, 3.01]33% [22%, 50%]67%80%——2026-08-01 → 2026-08-30
tmin00Znearest30d30 1 QC2.55 [1.81, 3.29]1.64+2.31 [1.27, 3.29]3.13 [2.20, 3.82]20% [7%, 37%]43%73%——2026-08-01 → 2026-08-30
tmin00Znearest90d30 1 QC2.55 [1.64, 3.28]1.64+2.31 [0.95, 3.23]3.13 [1.91, 3.79]20% [5%, 42%]43%73%——2026-08-01 → 2026-08-30
tmin00Znearest365d30 1 QC2.55 [1.61, 3.31]1.64+2.31 [0.94, 3.31]3.13 [1.93, 3.86]20% [4%, 45%]43%73%——2026-08-01 → 2026-08-30
tmin00Znearestall30 1 QC2.55 [1.60, 3.21]1.64+2.31 [0.89, 3.19]3.13 [1.86, 3.79]20% [5%, 42%]43%73%——2026-08-01 → 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 KATL, lead day 0, 00Z bilinear tmax+5+2.50-2.5-52026-08-01: -1.47 °F2026-08-02: -1.98 °F2026-08-03: -1.31 °F2026-08-04: +2.1 °F2026-08-05: -0.35 °F2026-08-06: +2.12 °F2026-08-07: +2.08 °F2026-08-08: +0.46 °F2026-08-09: +0.27 °F2026-08-10: -2.27 °F2026-08-11: +0.97 °F2026-08-12: -0.69 °F2026-08-13: +0.31 °F2026-08-14: +2.28 °F2026-08-15: -0.15 °F2026-08-16: +0.37 °F2026-08-17: +0.83 °F2026-08-18: +3 °F2026-08-19: -0.09 °F2026-08-20: +1.89 °F2026-08-21: -1.61 °F2026-08-22: -1.34 °F2026-08-23: -0.14 °F2026-08-24: +0.62 °F2026-08-25: +0.22 °F2026-08-26: +0.02 °F2026-08-27: +1.94 °F2026-08-28: -3.24 °F2026-08-29: -0.64 °F2026-08-30: +0.24 °F2026-08-012026-08-30°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 (30 days, newest first) · download CSV
Climatological dayError °F
2026-08-30+0.24
2026-08-29-0.64
2026-08-28-3.24
2026-08-27+1.94
2026-08-26+0.02
2026-08-25+0.22
2026-08-24+0.62
2026-08-23-0.14
2026-08-22-1.34
2026-08-21-1.61
2026-08-20+1.89
2026-08-19-0.09
2026-08-18+3.00
2026-08-17+0.83
2026-08-16+0.37
2026-08-15-0.15
2026-08-14+2.28
2026-08-13+0.31
2026-08-12-0.69
2026-08-11+0.97
2026-08-10-2.27
2026-08-09+0.27
2026-08-08+0.46
2026-08-07+2.08
2026-08-06+2.12
2026-08-05-0.35
2026-08-04+2.10
2026-08-03-1.31
2026-08-02-1.98
2026-08-01-1.47

Error distribution

Distribution of the signed daily error in °F for NCEP GFS at KATL, lead day 0-5 to -4.5 °F: 0 days-4.5 to -4 °F: 0 days-4 to -3.5 °F: 0 days-3.5 to -3 °F: 1 day-3 to -2.5 °F: 0 days-2.5 to -2 °F: 1 day-2 to -1.5 °F: 2 days-1.5 to -1 °F: 3 days-1 to -0.5 °F: 2 days-0.5 to 0 °F: 4 days0 to 0.5 °F: 7 days0.5 to 1 °F: 3 days1 to 1.5 °F: 0 days1.5 to 2 °F: 2 days2 to 2.5 °F: 4 days2.5 to 3 °F: 1 day3 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| 0.9 · P90 |e| 2.27 °F-50+5 °Fn = 30 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
-5.00-4.500
-4.50-4.000
-4.00-3.500
-3.50-3.001
-3.00-2.500
-2.50-2.001
-2.00-1.502
-1.50-1.003
-1.00-0.502
-0.50+0.004
+0.00+0.507
+0.50+1.003
+1.00+1.500
+1.50+2.002
+2.00+2.504
+2.50+3.001
+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
2026-08301.17 +0.15

Availability

ModelScored period DaysCoverage
NCEP GFS2026-08-01 → 2026-08-30 30
NCEP GFS2026-08-01 → 2026-08-30 30

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 30 -3.20 [-3.58, -2.74] yes
Aurora (GFS init) 2 -6.15 [-6.24, -6.06] yes
Aurora (IFS init) 3 -5.31 [-5.67, -4.91] yes
FourCastNet v2 (GFS init) 2 -3.23 [-3.32, -3.15] yes
FourCastNet v2 (IFS init) 3 -3.06 [-4.00, -2.46] yes
GraphCast (GFS init) 2 -4.60 [-5.24, -3.96] yes
GraphCast (IFS init) 28 -4.72 [-5.47, -3.94] yes
ECMWF IFS HRES 24 -1.60 [-2.29, -0.95] yes
Pangu-Weather (GFS init) 2 -2.06 [-2.46, -1.66] yes
Pangu-Weather (IFS init) 3 -1.91 [-2.12, -1.80] yes

Cite this page

ShortCastCheck, KATL gfs lead 0, methodology v0.2, accessed 2026-08-31, https://castcheck.zifanzhang.com/station/KATL/model/gfs/lead/0/

LongCastCheck (2026). Station-level verification of raw weather-model 2 m temperature forecasts: KATL (Atlanta Hartsfield), model gfs, lead day 0. Methodology version 0.2, schema version 0.1, data through 2026-08-30. Accessed 2026-08-31. https://castcheck.zifanzhang.com/station/KATL/model/gfs/lead/0/

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