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

All stations (mean of daily station errors) · GraphCast (IFS init) · lead day 7

ALL · graphcast_ifs · lead 7

Target day = initialization date + 7 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, 3 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
tmax00Zbilinear30d23 2 QC5.88 [5.34, 6.40]0.84-5.49 [-6.07, -4.91]7.04 [6.44, 7.59]7% [5%, 9%]15%25%-0.40+0.402026-08-08 → 2026-08-30
tmax00Zbilinear90d23 2 QC5.88 [5.30, 6.46]0.84-5.49 [-6.15, -4.88]7.04 [6.39, 7.62]7% [4%, 9%]15%25%-0.40+0.402026-08-08 → 2026-08-30
tmax00Zbilinear365d29 3 QC5.79 [5.30, 6.31]0.98-5.22 [-5.93, -4.51]6.97 [6.44, 7.50]8% [6%, 12%]17%27%-0.07+0.432026-01-08 → 2026-08-30
tmax00Zbilinearall29 3 QC5.79 [5.31, 6.34]0.98-5.22 [-5.97, -4.48]6.97 [6.45, 7.54]8% [6%, 12%]17%27%-0.07+0.432026-01-08 → 2026-08-30
tmax00Znearest30d23 2 QC6.00 [5.41, 6.54]0.84-5.44 [-6.04, -4.85]7.21 [6.53, 7.82]9% [7%, 10%]16%25%-0.43+0.402026-08-08 → 2026-08-30
tmax00Znearest90d23 2 QC6.00 [5.33, 6.62]0.84-5.44 [-6.12, -4.75]7.21 [6.43, 7.85]9% [6%, 11%]16%25%-0.43+0.402026-08-08 → 2026-08-30
tmax00Znearest365d29 3 QC5.89 [5.35, 6.45]0.98-5.19 [-5.90, -4.47]7.10 [6.49, 7.70]9% [7%, 10%]18%27%-0.09+0.432026-01-08 → 2026-08-30
tmax00Znearestall29 3 QC5.89 [5.37, 6.46]0.98-5.19 [-5.94, -4.46]7.10 [6.49, 7.73]9% [7%, 10%]18%27%-0.09+0.432026-01-08 → 2026-08-30
tmax12Zbilinear30d23 2 QC5.87 [5.22, 6.47]0.83-5.59 [-6.22, -4.93]6.89 [6.26, 7.43]6% [5%, 8%]14%23%-0.40+0.402026-08-08 → 2026-08-30
tmax12Zbilinear90d23 2 QC5.87 [5.19, 6.52]0.83-5.59 [-6.27, -4.91]6.89 [6.23, 7.48]6% [4%, 8%]14%23%-0.40+0.402026-08-08 → 2026-08-30
tmax12Zbilinear365d28 3 QC5.67 [4.97, 6.34]1.06-5.16 [-6.00, -4.03]6.78 [6.14, 7.37]8% [5%, 11%]17%26%-0.08+0.402026-01-08 → 2026-08-30
tmax12Zbilinearall28 3 QC5.67 [4.92, 6.37]1.06-5.16 [-6.08, -3.87]6.78 [6.13, 7.39]8% [5%, 12%]17%26%-0.08+0.402026-01-08 → 2026-08-30
tmax12Znearest30d23 2 QC5.98 [5.29, 6.62]0.85-5.54 [-6.18, -4.82]7.07 [6.35, 7.66]9% [6%, 11%]16%24%-0.43+0.392026-08-08 → 2026-08-30
tmax12Znearest90d23 2 QC5.98 [5.23, 6.71]0.85-5.54 [-6.25, -4.79]7.07 [6.31, 7.74]9% [6%, 11%]16%24%-0.43+0.392026-08-08 → 2026-08-30
tmax12Znearest365d28 3 QC5.77 [5.05, 6.49]1.06-5.12 [-5.98, -4.01]6.94 [6.21, 7.58]9% [7%, 11%]18%27%-0.10+0.402026-01-08 → 2026-08-30
tmax12Znearestall28 3 QC5.77 [5.00, 6.51]1.06-5.12 [-6.06, -3.86]6.94 [6.20, 7.61]9% [7%, 11%]18%27%-0.10+0.402026-01-08 → 2026-08-30
tmin00Zbilinear30d23 2 QC2.91 [2.74, 3.09]0.81+0.45 [-0.01, 0.93]3.74 [3.48, 3.99]24% [22%, 26%]44%60%+0.21+0.322026-08-08 → 2026-08-30
tmin00Zbilinear90d23 2 QC2.91 [2.72, 3.11]0.81+0.45 [-0.07, 1.07]3.74 [3.44, 4.02]24% [22%, 27%]44%60%+0.21+0.322026-08-08 → 2026-08-30
tmin00Zbilinear365d29 2 QC3.31 [2.81, 4.15]1.10+0.78 [0.15, 1.66]4.35 [3.60, 5.41]22% [17%, 25%]40%55%+0.34+0.382026-01-08 → 2026-08-30
tmin00Zbilinearall29 2 QC3.31 [2.81, 4.31]1.10+0.78 [0.13, 1.62]4.35 [3.59, 5.59]22% [15%, 26%]40%55%+0.34+0.382026-01-08 → 2026-08-30
tmin00Znearest30d23 2 QC2.95 [2.76, 3.16]0.82+0.48 [-0.00, 0.97]3.81 [3.53, 4.08]23% [19%, 26%]44%61%+0.20+0.312026-08-08 → 2026-08-30
tmin00Znearest90d23 2 QC2.95 [2.75, 3.18]0.82+0.48 [-0.06, 1.10]3.81 [3.51, 4.12]23% [19%, 26%]44%61%+0.20+0.312026-08-08 → 2026-08-30
tmin00Znearest365d29 2 QC3.38 [2.83, 4.28]1.11+0.80 [0.18, 1.69]4.52 [3.65, 5.72]21% [17%, 25%]41%57%+0.33+0.372026-01-08 → 2026-08-30
tmin00Znearestall29 2 QC3.38 [2.83, 4.44]1.11+0.80 [0.14, 1.67]4.52 [3.64, 5.92]21% [18%, 25%]41%57%+0.33+0.372026-01-08 → 2026-08-30
tmin12Zbilinear30d23 2 QC2.64 [2.44, 2.81]0.57+0.35 [0.05, 0.65]3.37 [3.09, 3.59]24% [21%, 27%]48%65%+0.28+0.522026-08-08 → 2026-08-30
tmin12Zbilinear90d23 2 QC2.64 [2.44, 2.83]0.57+0.35 [0.01, 0.69]3.37 [3.09, 3.61]24% [20%, 27%]48%65%+0.28+0.522026-08-08 → 2026-08-30
tmin12Zbilinear365d28 2 QC3.03 [2.52, 3.98]0.93+0.77 [0.17, 1.84]4.01 [3.20, 5.25]23% [18%, 27%]44%61%+0.39+0.462026-01-08 → 2026-08-30
tmin12Zbilinearall28 2 QC3.03 [2.51, 4.15]0.93+0.77 [0.12, 1.94]4.01 [3.19, 5.44]23% [17%, 26%]44%61%+0.39+0.462026-01-08 → 2026-08-30
tmin12Znearest30d23 2 QC2.69 [2.44, 2.88]0.59+0.38 [0.06, 0.67]3.45 [3.11, 3.69]23% [20%, 26%]49%66%+0.27+0.512026-08-08 → 2026-08-30
tmin12Znearest90d23 2 QC2.69 [2.43, 2.90]0.59+0.38 [0.02, 0.72]3.45 [3.11, 3.72]23% [20%, 26%]49%66%+0.27+0.512026-08-08 → 2026-08-30
tmin12Znearest365d28 2 QC3.10 [2.53, 4.12]0.94+0.80 [0.19, 1.87]4.19 [3.25, 5.59]22% [18%, 26%]45%61%+0.37+0.452026-01-08 → 2026-08-30
tmin12Znearestall28 2 QC3.10 [2.51, 4.29]0.94+0.80 [0.15, 1.96]4.19 [3.23, 5.78]22% [17%, 26%]45%61%+0.37+0.452026-01-08 → 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 ALL, lead day 7, 00Z bilinear tmax+10+50-5-102026-08-08: -5.25 °F2026-08-09: -6.12 °F2026-08-10: -3.43 °F2026-08-11: -5.11 °F2026-08-12: -5.57 °F2026-08-13: -3.47 °F2026-08-14: -5.21 °F2026-08-15: -5.7 °F2026-08-16: -3.3 °F2026-08-17: -4.94 °F2026-08-18: -6.13 °F2026-08-19: -6.63 °F2026-08-20: -4.83 °F2026-08-21: -5.48 °F2026-08-22: -4.81 °F2026-08-23: -7.05 °F2026-08-24: -8.86 °F2026-08-25: -6.25 °F2026-08-26: -5.24 °F2026-08-27: -5.25 °F2026-08-28: -6.5 °F2026-08-29: -5.69 °F2026-08-30: -5.37 °F2026-08-082026-08-30°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 (23 days, newest first) · download CSV
Climatological dayError °F
2026-08-30-5.37
2026-08-29-5.69
2026-08-28-6.50
2026-08-27-5.25
2026-08-26-5.24
2026-08-25-6.25
2026-08-24-8.86
2026-08-23-7.05
2026-08-22-4.81
2026-08-21-5.48
2026-08-20-4.83
2026-08-19-6.63
2026-08-18-6.13
2026-08-17-4.94
2026-08-16-3.30
2026-08-15-5.70
2026-08-14-5.21
2026-08-13-3.47
2026-08-12-5.57
2026-08-11-5.11
2026-08-10-3.43
2026-08-09-6.12
2026-08-08-5.25

Error distribution

Distribution of the signed daily error in °F for GraphCast (IFS init) at ALL, lead day 7-10 to -9 °F: 0 days-9 to -8 °F: 1 day-8 to -7 °F: 1 day-7 to -6 °F: 6 days-6 to -5 °F: 11 days-5 to -4 °F: 5 days-4 to -3 °F: 3 days-3 to -2 °F: 2 days-2 to -1 °F: 0 days-1 to 0 °F: 0 days0 to 1 °F: 0 days1 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| 5.25 · P90 |e| 6.53 °F-100+10 °Fn = 29 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
-9.00-8.001
-8.00-7.001
-7.00-6.006
-6.00-5.0011
-5.00-4.005
-4.00-3.003
-3.00-2.002
-2.00-1.000
-1.00+0.000
+0.00+1.000
+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
2026-08235.49 -5.49
2026-0164.20 -4.20

Availability

ModelScored period DaysCoverage
GraphCast (IFS init)2026-01-08 → 2026-08-30 29
GraphCast (IFS init)2026-01-08 → 2026-08-30 29
GraphCast (IFS init)2026-01-08 → 2026-08-30 28
GraphCast (IFS init)2026-01-08 → 2026-08-30 28

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 23 +1.50 [0.77, 1.97] yes
NCEP GFS 23 +1.96 [1.26, 2.60] yes
ECMWF IFS HRES 18 +2.47 [2.06, 2.83] yes
Persistence (baseline) 23 +1.69 [0.89, 2.42] yes

Cite this page

ShortCastCheck, ALL graphcast_ifs lead 7, methodology v0.2, accessed 2026-08-31, https://castcheck.zifanzhang.com/station/ALL/model/graphcast_ifs/lead/7/

LongCastCheck (2026). Station-level verification of raw weather-model 2 m temperature forecasts: ALL (All stations (mean of daily station errors)), model graphcast_ifs, lead day 7. Methodology version 0.2, schema version 0.1, data through 2026-08-30. Accessed 2026-08-31. https://castcheck.zifanzhang.com/station/ALL/model/graphcast_ifs/lead/7/

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