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 · GraphCast (IFS init) · lead day 0

KATL · graphcast_ifs · 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, GRAP_v100, starting 2026-01-01 — 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
tmax00Zbilinear30d28 1 QC5.90 [5.16, 6.55]1.30-5.90 [-6.55, -5.16]6.11 [5.39, 6.70]0% [0%, 0%]0%0%——2026-08-01 → 2026-08-30
tmax00Zbilinear90d28 1 QC5.90 [5.11, 6.63]1.30-5.90 [-6.63, -5.11]6.11 [5.33, 6.79]0% [0%, 0%]0%0%——2026-08-01 → 2026-08-30
tmax00Zbilinear365d34 1 QC5.67 [4.87, 6.48]1.47-5.65 [-6.48, -4.81]5.98 [5.27, 6.64]3% [0%, 11%]3%3%——2026-01-01 → 2026-08-30
tmax00Zbilinearall34 1 QC5.67 [4.88, 6.45]1.47-5.65 [-6.45, -4.82]5.98 [5.23, 6.62]3% [0%, 11%]3%3%——2026-01-01 → 2026-08-30
tmax00Znearest30d28 1 QC5.79 [4.98, 6.52]1.29-5.79 [-6.52, -4.98]6.01 [5.22, 6.68]0% [0%, 0%]0%7%——2026-08-01 → 2026-08-30
tmax00Znearest90d28 1 QC5.79 [4.94, 6.61]1.29-5.79 [-6.61, -4.94]6.01 [5.14, 6.76]0% [0%, 0%]0%7%——2026-08-01 → 2026-08-30
tmax00Znearest365d34 1 QC5.60 [4.78, 6.40]1.50-5.58 [-6.40, -4.74]5.91 [5.16, 6.59]3% [0%, 11%]3%9%——2026-01-01 → 2026-08-30
tmax00Znearestall34 1 QC5.60 [4.79, 6.40]1.50-5.58 [-6.40, -4.74]5.91 [5.11, 6.57]3% [0%, 11%]3%9%——2026-01-01 → 2026-08-30
tmin00Zbilinear30d28 1 QC1.54 [1.14, 1.99]1.36-0.83 [-1.58, -0.22]1.94 [1.43, 2.35]36% [21%, 47%]71%86%——2026-08-01 → 2026-08-30
tmin00Zbilinear90d28 1 QC1.54 [1.08, 2.06]1.36-0.83 [-1.82, -0.24]1.94 [1.33, 2.38]36% [17%, 50%]71%86%——2026-08-01 → 2026-08-30
tmin00Zbilinear365d34 1 QC1.50 [1.13, 1.97]1.42-0.55 [-1.34, 0.16]1.89 [1.42, 2.33]38% [22%, 51%]71%88%——2026-01-01 → 2026-08-30
tmin00Zbilinearall34 1 QC1.50 [1.13, 1.99]1.42-0.55 [-1.41, 0.09]1.89 [1.40, 2.35]38% [22%, 50%]71%88%——2026-01-01 → 2026-08-30
tmin00Znearest30d28 1 QC1.58 [1.18, 2.02]1.39-0.80 [-1.58, -0.15]1.97 [1.44, 2.41]39% [23%, 54%]68%86%——2026-08-01 → 2026-08-30
tmin00Znearest90d28 1 QC1.58 [1.13, 2.09]1.39-0.80 [-1.80, -0.18]1.97 [1.37, 2.40]39% [17%, 58%]68%86%——2026-08-01 → 2026-08-30
tmin00Znearest365d34 1 QC1.54 [1.18, 1.97]1.44-0.56 [-1.31, 0.14]1.92 [1.46, 2.35]41% [23%, 57%]65%88%——2026-01-01 → 2026-08-30
tmin00Znearestall34 1 QC1.54 [1.18, 1.99]1.44-0.56 [-1.40, 0.07]1.92 [1.44, 2.34]41% [23%, 56%]65%88%——2026-01-01 → 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 GraphCast (IFS init) at KATL, lead day 0, 00Z bilinear tmax+10+50-5-102026-08-01: -4.16 °F2026-08-02: -8.55 °F2026-08-03: -5.7 °F2026-08-04: -4.62 °F2026-08-05: -7.89 °F2026-08-06: -8.27 °F2026-08-07: -5.24 °F2026-08-08: -5.61 °F2026-08-09: -5.49 °F2026-08-10: -8.43 °F2026-08-11: -8.42 °F2026-08-12: -6.28 °F2026-08-13: -7.33 °F2026-08-14: -6.21 °F2026-08-15: -6.23 °F2026-08-16: -5.6 °F2026-08-17: -3.02 °F2026-08-18: -3.45 °F2026-08-19: -4.02 °F2026-08-20: -4.48 °F2026-08-21: -5.61 °F2026-08-22: -8.17 °F2026-08-23: -3.69 °F2026-08-24: -4.28 °F2026-08-25: -5.85 °F2026-08-28: -7.33 °F2026-08-29: -5.63 °F2026-08-30: -5.58 °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
2026-08-30-5.58
2026-08-29-5.63
2026-08-28-7.33
2026-08-27—
2026-08-26—
2026-08-25-5.85
2026-08-24-4.28
2026-08-23-3.69
2026-08-22-8.17
2026-08-21-5.61
2026-08-20-4.48
2026-08-19-4.02
2026-08-18-3.45
2026-08-17-3.02
2026-08-16-5.60
2026-08-15-6.23
2026-08-14-6.21
2026-08-13-7.33
2026-08-12-6.28
2026-08-11-8.42
2026-08-10-8.43
2026-08-09-5.49
2026-08-08-5.61
2026-08-07-5.24
2026-08-06-8.27
2026-08-05-7.89
2026-08-04-4.62
2026-08-03-5.70
2026-08-02-8.55
2026-08-01-4.16

Error distribution

Distribution of the signed daily error in °F for GraphCast (IFS init) at KATL, lead day 0-10 to -9 °F: 0 days-9 to -8 °F: 5 days-8 to -7 °F: 5 days-7 to -6 °F: 3 days-6 to -5 °F: 10 days-5 to -4 °F: 5 days-4 to -3 °F: 5 days-3 to -2 °F: 0 days-2 to -1 °F: 0 days-1 to 0 °F: 0 days0 to 1 °F: 1 day1 to 2 °F: 0 days2 to 3 °F: 0 days3 to 4 °F: 0 days4 to 5 °F: 0 days5 to 6 °F: 0 days6 to 7 °F: 0 days7 to 8 °F: 0 days8 to 9 °F: 0 days9 to 10 °F: 0 daysdashed: P50 |e| 5.62 · P90 |e| 8.24 °F-100+10 °Fn = 34 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
-10.00-9.000
-9.00-8.005
-8.00-7.005
-7.00-6.003
-6.00-5.0010
-5.00-4.005
-4.00-3.005
-3.00-2.000
-2.00-1.000
-1.00+0.000
+0.00+1.001
+1.00+2.000
+2.00+3.000
+3.00+4.000
+4.00+5.000
+5.00+6.000
+6.00+7.000
+7.00+8.000
+8.00+9.000
+9.00+10.000

Month by month

00Z bilinear daily maximum, newest first.

Monthn MAE °FBias °F
2026-08285.90 -5.90
2026-0164.62 -4.50

Availability

ModelScored period DaysCoverage
GraphCast (IFS init)2026-01-01 → 2026-08-30 34
GraphCast (IFS init)2026-01-01 → 2026-08-30 34

Compared with other models

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

Other modelCommon days MAE difference °FDistinguishable
ECMWF AIFS Single 28 +1.56 [0.79, 2.30] yes
Aurora (GFS init) 2 -0.98 [-1.07, -0.90] yes
Aurora (IFS init) 3 -0.51 [-0.82, -0.33] yes
FourCastNet v2 (GFS init) 2 +1.93 [1.67, 2.20] yes
FourCastNet v2 (IFS init) 3 +1.75 [0.08, 2.88] yes
NCEP GFS 28 +4.72 [3.94, 5.47] yes
GraphCast (GFS init) 2 +0.57 [0.11, 1.03] yes
ECMWF IFS HRES 22 +3.13 [2.07, 4.20] yes
Pangu-Weather (GFS init) 2 +3.11 [2.88, 3.33] yes
Pangu-Weather (IFS init) 3 +2.90 [2.20, 3.25] yes

Cite this page

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

LongCastCheck (2026). Station-level verification of raw weather-model 2 m temperature forecasts: KATL (Atlanta Hartsfield), model graphcast_ifs, 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/graphcast_ifs/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