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 · GraphCast (IFS init) · lead day 0

KMIA · 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, 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
tmax00Zbilinear30d28 2 QC3.66 [3.11, 4.29]1.48-3.66 [-4.29, -3.11]4.08 [3.39, 4.75]4% [0%, 10%]14%50%——2026-08-01 → 2026-08-30
tmax00Zbilinear90d28 2 QC3.66 [3.06, 4.59]1.48-3.66 [-4.59, -3.06]4.08 [3.38, 5.00]4% [0%, 10%]14%50%——2026-08-01 → 2026-08-30
tmax00Zbilinear365d34 2 QC3.48 [2.90, 4.22]1.35-3.48 [-4.22, -2.90]3.88 [3.13, 4.68]3% [0%, 9%]18%53%——2026-01-01 → 2026-08-30
tmax00Zbilinearall34 2 QC3.48 [2.91, 4.22]1.35-3.48 [-4.22, -2.91]3.88 [3.15, 4.64]3% [0%, 9%]18%53%——2026-01-01 → 2026-08-30
tmax00Znearest30d28 2 QC4.06 [3.64, 4.50]1.30-4.06 [-4.50, -3.64]4.38 [3.87, 4.86]0% [0%, 0%]11%18%——2026-08-01 → 2026-08-30
tmax00Znearest90d28 2 QC4.06 [3.59, 4.71]1.30-4.06 [-4.71, -3.59]4.38 [3.84, 5.06]0% [0%, 0%]11%18%——2026-08-01 → 2026-08-30
tmax00Znearest365d34 2 QC3.85 [3.32, 4.42]1.23-3.85 [-4.42, -3.32]4.17 [3.53, 4.79]0% [0%, 0%]9%26%——2026-01-01 → 2026-08-30
tmax00Znearestall34 2 QC3.85 [3.31, 4.39]1.23-3.85 [-4.39, -3.31]4.17 [3.57, 4.74]0% [0%, 0%]9%26%——2026-01-01 → 2026-08-30
tmin00Zbilinear30d28 2 QC1.47 [0.99, 2.11]1.50+0.66 [-0.06, 1.67]2.10 [1.25, 2.95]54% [43%, 65%]82%86%——2026-08-01 → 2026-08-30
tmin00Zbilinear90d28 2 QC1.47 [0.98, 2.31]1.50+0.66 [-0.09, 1.92]2.10 [1.23, 3.18]54% [40%, 68%]82%86%——2026-08-01 → 2026-08-30
tmin00Zbilinear365d34 2 QC1.54 [1.08, 2.20]1.56+0.78 [0.01, 1.71]2.14 [1.38, 3.01]50% [37%, 62%]79%82%——2026-01-01 → 2026-08-30
tmin00Zbilinearall34 2 QC1.54 [1.07, 2.28]1.56+0.78 [0.02, 1.97]2.14 [1.35, 3.10]50% [38%, 64%]79%82%——2026-01-01 → 2026-08-30
tmin00Znearest30d28 2 QC1.70 [1.15, 2.51]1.43+1.41 [0.70, 2.45]2.38 [1.45, 3.38]46% [28%, 63%]75%82%——2026-08-01 → 2026-08-30
tmin00Znearest90d28 2 QC1.70 [1.11, 2.73]1.43+1.41 [0.66, 2.64]2.38 [1.43, 3.62]46% [23%, 64%]75%82%——2026-08-01 → 2026-08-30
tmin00Znearest365d34 2 QC1.83 [1.18, 2.63]1.52+1.58 [0.80, 2.53]2.53 [1.57, 3.46]44% [25%, 60%]71%79%——2026-01-01 → 2026-08-30
tmin00Znearestall34 2 QC1.83 [1.19, 2.84]1.52+1.58 [0.78, 2.78]2.53 [1.57, 3.65]44% [25%, 61%]71%79%——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 KMIA, lead day 0, 00Z bilinear tmax+10+50-5-102026-08-01: -2.7 °F2026-08-02: -1.32 °F2026-08-03: -7.05 °F2026-08-04: -4.56 °F2026-08-05: -5.52 °F2026-08-06: -4.27 °F2026-08-07: -2.99 °F2026-08-08: -1.1 °F2026-08-09: -2.79 °F2026-08-10: -1.74 °F2026-08-11: -3 °F2026-08-12: -2.55 °F2026-08-13: -4.87 °F2026-08-14: -4.38 °F2026-08-15: -0.9 °F2026-08-16: -2.7 °F2026-08-17: -2.71 °F2026-08-18: -4.82 °F2026-08-19: -2.69 °F2026-08-20: -2.05 °F2026-08-21: -7.68 °F2026-08-22: -3.07 °F2026-08-23: -3.01 °F2026-08-24: -3.29 °F2026-08-25: -2.9 °F2026-08-28: -4.64 °F2026-08-29: -8.08 °F2026-08-30: -5.06 °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
2026-08-30-5.06
2026-08-29-8.08
2026-08-28-4.64
2026-08-27—
2026-08-26—
2026-08-25-2.90
2026-08-24-3.29
2026-08-23-3.01
2026-08-22-3.07
2026-08-21-7.68
2026-08-20-2.05
2026-08-19-2.69
2026-08-18-4.82
2026-08-17-2.71
2026-08-16-2.70
2026-08-15-0.90
2026-08-14-4.38
2026-08-13-4.87
2026-08-12-2.55
2026-08-11-3.00
2026-08-10-1.74
2026-08-09-2.79
2026-08-08-1.10
2026-08-07-2.99
2026-08-06-4.27
2026-08-05-5.52
2026-08-04-4.56
2026-08-03-7.05
2026-08-02-1.32
2026-08-01-2.70

Error distribution

Distribution of the signed daily error in °F for GraphCast (IFS init) at KMIA, lead day 0-10 to -9 °F: 0 days-9 to -8 °F: 1 day-8 to -7 °F: 2 days-7 to -6 °F: 0 days-6 to -5 °F: 2 days-5 to -4 °F: 6 days-4 to -3 °F: 5 days-3 to -2 °F: 12 days-2 to -1 °F: 5 days-1 to 0 °F: 1 day0 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| 2.99 · P90 |e| 5.38 °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
-9.00-8.001
-8.00-7.002
-7.00-6.000
-6.00-5.002
-5.00-4.006
-4.00-3.005
-3.00-2.0012
-2.00-1.005
-1.00+0.001
+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
2026-08283.66 -3.66
2026-0162.66 -2.66

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 -0.29 [-0.72, 0.27] no
Aurora (GFS init) 2 -0.39 [-2.39, 1.61] no
Aurora (IFS init) 3 -1.01 [-2.03, -0.50] yes
FourCastNet v2 (GFS init) 2 +2.03 [0.10, 3.95] yes
FourCastNet v2 (IFS init) 3 +0.84 [-0.93, 1.73] no
NCEP GFS 28 +2.22 [1.34, 3.40] yes
GraphCast (GFS init) 2 +0.52 [-1.45, 2.49] no
ECMWF IFS HRES 22 +0.39 [-0.79, 1.29] no
Pangu-Weather (GFS init) 2 +0.74 [-1.34, 2.82] no
Pangu-Weather (IFS init) 3 +1.28 [0.18, 1.82] yes

Cite this page

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

LongCastCheck (2026). Station-level verification of raw weather-model 2 m temperature forecasts: KMIA (Miami Intl), 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/KMIA/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