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 2

ALL · graphcast_ifs · lead 2

Target day = initialization date + 2 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
tmax00Zbilinear30d26 1 QC5.09 [4.85, 5.32]0.43-4.97 [-5.23, -4.72]5.66 [5.38, 5.88]6% [4%, 7%]11%22%-0.53+0.612026-08-03 → 2026-08-30
tmax00Zbilinear90d26 1 QC5.09 [4.82, 5.33]0.43-4.97 [-5.23, -4.69]5.66 [5.33, 5.89]6% [4%, 7%]11%22%-0.53+0.612026-08-03 → 2026-08-30
tmax00Zbilinear365d32 1 QC5.04 [4.81, 5.27]0.45-4.93 [-5.17, -4.69]5.67 [5.39, 5.86]6% [5%, 7%]12%23%-0.28+0.702026-01-03 → 2026-08-30
tmax00Zbilinearall32 1 QC5.04 [4.82, 5.27]0.45-4.93 [-5.18, -4.71]5.67 [5.41, 5.87]6% [5%, 7%]12%23%-0.28+0.702026-01-03 → 2026-08-30
tmax00Znearest30d26 1 QC5.27 [5.04, 5.54]0.47-4.92 [-5.25, -4.59]5.89 [5.60, 6.17]7% [6%, 8%]13%21%-0.59+0.582026-08-03 → 2026-08-30
tmax00Znearest90d26 1 QC5.27 [5.02, 5.55]0.47-4.92 [-5.25, -4.55]5.89 [5.53, 6.20]7% [5%, 8%]13%21%-0.59+0.582026-08-03 → 2026-08-30
tmax00Znearest365d32 1 QC5.21 [4.98, 5.47]0.48-4.90 [-5.18, -4.60]5.88 [5.61, 6.14]7% [5%, 8%]14%23%-0.32+0.682026-01-03 → 2026-08-30
tmax00Znearestall32 1 QC5.21 [4.98, 5.48]0.48-4.90 [-5.20, -4.61]5.88 [5.61, 6.14]7% [5%, 8%]14%23%-0.32+0.682026-01-03 → 2026-08-30
tmax12Zbilinear30d26 1 QC5.42 [5.14, 5.69]0.44-5.35 [-5.63, -5.04]6.04 [5.71, 6.29]4% [2%, 6%]12%20%-0.64+0.612026-08-03 → 2026-08-30
tmax12Zbilinear90d26 1 QC5.42 [5.11, 5.71]0.44-5.35 [-5.65, -5.01]6.04 [5.67, 6.31]4% [2%, 6%]12%20%-0.64+0.612026-08-03 → 2026-08-30
tmax12Zbilinear365d31 1 QC5.36 [5.10, 5.64]0.48-5.29 [-5.58, -5.00]6.04 [5.73, 6.28]5% [3%, 6%]13%22%-0.38+0.682026-01-03 → 2026-08-30
tmax12Zbilinearall31 1 QC5.36 [5.08, 5.64]0.48-5.29 [-5.58, -5.00]6.04 [5.75, 6.29]5% [3%, 6%]13%22%-0.38+0.682026-01-03 → 2026-08-30
tmax12Znearest30d26 1 QC5.60 [5.33, 5.90]0.48-5.27 [-5.61, -4.89]6.24 [5.89, 6.56]5% [3%, 7%]12%19%-0.69+0.582026-08-03 → 2026-08-30
tmax12Znearest90d26 1 QC5.60 [5.29, 5.92]0.48-5.27 [-5.63, -4.83]6.24 [5.82, 6.59]5% [3%, 7%]12%19%-0.69+0.582026-08-03 → 2026-08-30
tmax12Znearest365d31 1 QC5.53 [5.23, 5.84]0.52-5.23 [-5.57, -4.88]6.23 [5.91, 6.53]5% [3%, 7%]13%21%-0.43+0.662026-01-03 → 2026-08-30
tmax12Znearestall31 1 QC5.53 [5.23, 5.84]0.52-5.23 [-5.56, -4.88]6.23 [5.91, 6.55]5% [3%, 7%]13%21%-0.43+0.662026-01-03 → 2026-08-30
tmin00Zbilinear30d26 1 QC2.17 [2.01, 2.34]0.56+0.58 [0.30, 0.87]2.86 [2.58, 3.10]31% [29%, 33%]56%74%+0.25+0.362026-08-03 → 2026-08-30
tmin00Zbilinear90d26 1 QC2.17 [2.02, 2.37]0.56+0.58 [0.21, 0.89]2.86 [2.60, 3.15]31% [29%, 33%]56%74%+0.25+0.362026-08-03 → 2026-08-30
tmin00Zbilinear365d32 1 QC2.34 [2.09, 2.75]0.64+0.73 [0.40, 1.10]3.18 [2.73, 3.80]29% [25%, 32%]54%72%+0.32+0.522026-01-03 → 2026-08-30
tmin00Zbilinearall32 1 QC2.34 [2.07, 2.75]0.64+0.73 [0.32, 1.10]3.18 [2.70, 3.83]29% [26%, 32%]54%72%+0.32+0.522026-01-03 → 2026-08-30
tmin00Znearest30d26 1 QC2.19 [1.99, 2.37]0.57+0.60 [0.30, 0.86]2.92 [2.61, 3.17]32% [28%, 36%]57%73%+0.24+0.362026-08-03 → 2026-08-30
tmin00Znearest90d26 1 QC2.19 [2.00, 2.40]0.57+0.60 [0.21, 0.90]2.92 [2.63, 3.23]32% [27%, 36%]57%73%+0.24+0.362026-08-03 → 2026-08-30
tmin00Znearest365d32 1 QC2.39 [2.08, 2.86]0.63+0.72 [0.39, 1.06]3.24 [2.78, 3.88]30% [23%, 35%]54%71%+0.31+0.532026-01-03 → 2026-08-30
tmin00Znearestall32 1 QC2.39 [2.06, 2.87]0.63+0.72 [0.33, 1.07]3.24 [2.74, 3.89]30% [23%, 34%]54%71%+0.31+0.532026-01-03 → 2026-08-30
tmin12Zbilinear30d26 1 QC2.08 [1.94, 2.20]0.54+0.31 [0.01, 0.57]2.73 [2.51, 2.91]34% [32%, 35%]58%75%+0.29+0.382026-08-03 → 2026-08-30
tmin12Zbilinear90d26 1 QC2.08 [1.95, 2.23]0.54+0.31 [-0.06, 0.59]2.73 [2.52, 2.96]34% [32%, 35%]58%75%+0.29+0.382026-08-03 → 2026-08-30
tmin12Zbilinear365d31 1 QC2.20 [1.99, 2.53]0.65+0.46 [0.11, 0.85]2.98 [2.60, 3.52]32% [29%, 35%]58%74%+0.36+0.502026-01-03 → 2026-08-30
tmin12Zbilinearall31 1 QC2.20 [1.99, 2.51]0.65+0.46 [0.04, 0.84]2.98 [2.60, 3.50]32% [29%, 35%]58%74%+0.36+0.502026-01-03 → 2026-08-30
tmin12Znearest30d26 1 QC2.13 [1.94, 2.27]0.54+0.35 [0.03, 0.60]2.81 [2.55, 3.00]34% [30%, 38%]58%75%+0.27+0.392026-08-03 → 2026-08-30
tmin12Znearest90d26 1 QC2.13 [1.96, 2.30]0.54+0.35 [-0.05, 0.63]2.81 [2.57, 3.04]34% [28%, 37%]58%75%+0.27+0.392026-08-03 → 2026-08-30
tmin12Znearest365d31 1 QC2.26 [2.01, 2.62]0.64+0.48 [0.12, 0.83]3.05 [2.66, 3.57]32% [25%, 36%]57%74%+0.34+0.512026-01-03 → 2026-08-30
tmin12Znearestall31 1 QC2.26 [2.01, 2.60]0.64+0.48 [0.07, 0.83]3.05 [2.66, 3.56]32% [26%, 36%]57%74%+0.34+0.512026-01-03 → 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 2, 00Z bilinear tmax+10+50-5-102026-08-03: -4.8 °F2026-08-04: -5.07 °F2026-08-05: -5.23 °F2026-08-06: -4.38 °F2026-08-07: -4.75 °F2026-08-08: -4.31 °F2026-08-09: -3.96 °F2026-08-10: -4.31 °F2026-08-11: -4.29 °F2026-08-12: -6.12 °F2026-08-13: -4.61 °F2026-08-14: -5.43 °F2026-08-15: -4.93 °F2026-08-16: -4.52 °F2026-08-17: -5.24 °F2026-08-18: -4.65 °F2026-08-19: -5.09 °F2026-08-20: -5.39 °F2026-08-21: -5.13 °F2026-08-22: -5.55 °F2026-08-23: -5.51 °F2026-08-24: -5.81 °F2026-08-25: -4.56 °F2026-08-26: -5.45 °F2026-08-27: -5.2 °F2026-08-30: -5.06 °F2026-08-032026-08-30°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 (28 days, newest first) · download CSV
Climatological dayError °F
2026-08-30-5.06
2026-08-29—
2026-08-28—
2026-08-27-5.20
2026-08-26-5.45
2026-08-25-4.56
2026-08-24-5.81
2026-08-23-5.51
2026-08-22-5.55
2026-08-21-5.13
2026-08-20-5.39
2026-08-19-5.09
2026-08-18-4.65
2026-08-17-5.24
2026-08-16-4.52
2026-08-15-4.93
2026-08-14-5.43
2026-08-13-4.61
2026-08-12-6.12
2026-08-11-4.29
2026-08-10-4.31
2026-08-09-3.96
2026-08-08-4.31
2026-08-07-4.75
2026-08-06-4.38
2026-08-05-5.23
2026-08-04-5.07
2026-08-03-4.80

Error distribution

Distribution of the signed daily error in °F for GraphCast (IFS init) at ALL, lead day 2-10 to -9 °F: 0 days-9 to -8 °F: 0 days-8 to -7 °F: 0 days-7 to -6 °F: 1 day-6 to -5 °F: 15 days-5 to -4 °F: 14 days-4 to -3 °F: 2 days-3 to -2 °F: 0 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| 4.99 · P90 |e| 5.55 °F-100+10 °Fn = 32 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
-9.00-8.000
-8.00-7.000
-7.00-6.001
-6.00-5.0015
-5.00-4.0014
-4.00-3.002
-3.00-2.000
-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
2026-08264.97 -4.97
2026-0164.76 -4.76

Availability

ModelScored period DaysCoverage
GraphCast (IFS init)2026-01-03 → 2026-08-30 32
GraphCast (IFS init)2026-01-03 → 2026-08-30 32
GraphCast (IFS init)2026-01-03 → 2026-08-30 31
GraphCast (IFS init)2026-01-03 → 2026-08-30 31

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 26 +1.50 [1.31, 1.69] yes
Aurora (IFS init) 1 +0.08 [0.08, 0.08] yes
FourCastNet v2 (IFS init) 1 +1.44 [1.44, 1.44] yes
NCEP GFS 26 +2.02 [1.74, 2.20] yes
ECMWF IFS HRES 21 +1.68 [1.35, 1.99] yes
Pangu-Weather (IFS init) 1 +1.06 [1.06, 1.06] yes
Persistence (baseline) 26 +1.77 [1.19, 2.13] yes

Cite this page

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

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 2. 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/2/

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