The 30 DWD stations scored on this page. Hover over or tap a point to see the site.
Results at the German stations
Show
Version
Reference
Forecasts
AIFS 2: in service from 12 May 2026; forecasts issued from 13 May 2026, scored for valid days up to 12 November 2026.
2 m temperature: skill against climatology
Raw forecasts, 174 days, 23 May 2026 to 12 November 2026. Hover for values; AIFS shows its 95% range, others on hover.
2 m temperature: skill against persistence
Raw forecasts, 174 days, 23 May 2026 to 12 November 2026, on the 10,241 of 10,401 station-times where persistence can be scored as well. Hover for values; AIFS shows its 95% range, others on hover.
2 m temperature: skill against climatology
Bias-corrected forecasts, 144 days, 22 June 2026 to 12 November 2026. Hover for values; AIFS shows its 95% range, others on hover.
2 m temperature: skill against persistence
Bias-corrected forecasts, 144 days, 22 June 2026 to 12 November 2026, on the 8,456 of 8,602 station-times where persistence can be scored as well. Hover for values; AIFS shows its 95% range, others on hover.
AIFS's RMSE against each model, in %, over 174 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
▼−10%, AIFS lower error, tested
▼−13%, AIFS lower error, tested
▼−13%, AIFS lower error, tested
−9%, not separated
−11%, not separated
GFS
▼−29%, AIFS lower error, tested
▼−29%, AIFS lower error, tested
▼−24%, AIFS lower error, tested
−20%, not separated
−15%, not separated
AIFS's RMSE against each model, in %, over 144 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
−13%, not separated
−16%, not separated
−14%, not separated
−7%, not separated
−13%, not separated
GFS
−30%, not separated
−28%, not separated
−22%, not separated
−20%, not separated
−16%, not separated
10 m wind speed: skill against climatology
Raw forecasts, 174 days, 23 May 2026 to 12 November 2026. Hover for values; AIFS shows its 95% range, others on hover.
10 m wind speed: skill against persistence
Raw forecasts, 174 days, 23 May 2026 to 12 November 2026, on the 10,081 of 10,351 station-times where persistence can be scored as well. Hover for values; AIFS shows its 95% range, others on hover.
10 m wind speed: skill against climatology
Bias-corrected forecasts, 144 days, 22 June 2026 to 12 November 2026. Hover for values; AIFS shows its 95% range, others on hover.
10 m wind speed: skill against persistence
Bias-corrected forecasts, 144 days, 22 June 2026 to 12 November 2026, on the 8,247 of 8,495 station-times where persistence can be scored as well. Hover for values; AIFS shows its 95% range, others on hover.
AIFS's RMSE against each model, in %, over 174 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
−1%, not separated
−3%, not separated
▼−7%, AIFS lower error, tested
−5%, not separated
0%, not separated
GFS
−4%, not separated
▼−8%, AIFS lower error, tested
▼−10%, AIFS lower error, tested
▼−10%, AIFS lower error, tested
−3%, not separated
AIFS's RMSE against each model, in %, over 144 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
−1%, not separated
−6%, not separated
−9%, not separated
−7%, not separated
−1%, not separated
GFS
−10%, not separated
−14%, not separated
−14%, not separated
−15%, not separated
−5%, not separated
Solar energy: skill against climatology
Raw forecasts, 174 days, 23 May 2026 to 12 November 2026. Hover for values; AIFS shows its 95% range, others on hover.
Solar energy: skill against persistence
Raw forecasts, 174 days, 23 May 2026 to 12 November 2026, on the 4,708 of 5,123 station-times where persistence can be scored as well. Hover for values; AIFS shows its 95% range, others on hover.
Solar energy: skill against climatology
Bias-corrected forecasts, 143 days, 23 June 2026 to 12 November 2026. Hover for values; AIFS shows its 95% range, others on hover.
Solar energy: skill against persistence
Bias-corrected forecasts, 142 days, 24 June 2026 to 12 November 2026: those of the 143 days scored on which persistence can be scored as well. Hover for values; AIFS shows its 95% range, others on hover.
AIFS's RMSE against each model, in %, over 174 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
▼−15%, AIFS lower error, tested
▼−16%, AIFS lower error, tested
▼−13%, AIFS lower error, tested
–
–
GFS
▼−23%, AIFS lower error, tested
▼−29%, AIFS lower error, tested
▼−25%, AIFS lower error, tested
▼−19%, AIFS lower error, tested
−13%, not separated
ICON is not compared on solar: its forecasts come from Open-Meteo, which assembles the day differently.
AIFS's RMSE against each model, in %, over 143 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
−11%, not separated
−12%, not separated
−11%, not separated
–
–
GFS
−22%, not separated
−24%, not separated
−21%, not separated
−15%, not separated
−10%, not separated
ICON is not compared on solar: its forecasts come from Open-Meteo, which assembles the day differently.
▼AIFS lower error, tested (better)
▲AIFS higher error, tested (worse)
Not separated: no clear difference, too few days, or the verdict depends on how the station is read or on the height correction (hover says which)
Each number is AIFS's RMSE minus the other model's, as a percentage of the other model's RMSE: minus means AIFS's was lower (better). Hover or tap a cell for the difference in units, its 95% interval and the p-value.
Skill against climatology at day 7, 2 m temperature: AIFS 29%, IFS 22%, GFS 11%. Over these days, AIFS had the highest skill against climatology at this lead. The climatology is not adjusted for warming and ran 0.9 K too cold over these days, which raises every model's temperature skill against it.
Skill against climatology at day 7, 10 m wind: AIFS −5%, IFS −11%, GFS −16%. Over these days, no model had a lower RMSE than climatology at this lead, and AIFS came closest to it.
Skill against climatology at day 7, daily solar energy: AIFS 16%, GFS −4%. Over these days, AIFS had the highest skill against climatology at this lead.
Bias, activity and eventsWhether a model runs high or low, how much it varies, and how it does in strong wind and frost
What an average error hides: whether a model runs high or low, whether it varies as much as the weather, and how it does when it matters. Bold and underlined marks the value nearest the ideal at each lead, as printed. It is not a verdict: these values are not tested against each other. Hover an event score for its 95% interval. The thresholds were fixed before scoring.
2 m temperature: bias, in °CForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
+0.08
+0.08
+0.21
+0.43
−0.15
IFS
+0.11
+0.14
+0.36
+0.53
+0.47
GFS
−0.14
−0.08
+0.02
−0.02
−0.21
2 m temperature: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.98
0.97
0.99
0.98
0.95
IFS
0.95
0.94
0.97
0.96
0.99
GFS
0.94
0.94
0.94
0.96
0.98
2 m temperature of 0 °C or belowObserved 20 times on 7 days in 10401 station-times (0.2%). Too few events to score: at least 30, on at least 10 days, are needed.
2 m temperature: bias, in °CForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
−0.07
−0.07
−0.13
−0.27
−0.14
IFS
−0.01
+0.02
−0.01
−0.07
−0.19
GFS
+0.17
+0.21
+0.25
+0.12
+0.05
2 m temperature: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.95
0.94
0.96
0.94
0.94
IFS
0.93
0.94
0.98
1.00
0.98
GFS
0.95
0.97
0.97
0.99
0.98
2 m temperature of 0 °C or belowObserved 20 times on 7 days in 8602 station-times (0.2%). Too few events to score: at least 30, on at least 10 days, are needed.
10 m wind: bias, in m/sForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
−0.26
−0.34
−0.38
−0.45
−0.42
IFS
−0.21
−0.26
−0.27
−0.34
−0.38
GFS
+0.02
−0.06
−0.07
−0.06
−0.08
10 m wind: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.91
0.89
0.85
0.83
0.87
IFS
0.95
0.94
0.92
0.89
0.86
GFS
1.00
0.98
0.93
0.95
0.95
10 m wind of 8 m/s or more: frequency biasObserved 285 times on 78 days in 10351 station-times (3%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.47
0.38
0.34
0.26
0.36
IFS
0.62
0.64
0.57
0.49
0.44
GFS
1.09
1.01
0.88
0.92
0.85
10 m wind of 8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.31
0.23
0.19
0.12
0.05
IFS
0.35
0.31
0.20
0.14
0.07
GFS
0.36
0.27
0.21
0.14
0.10
10 m wind of 10.8 m/s or more: frequency biasObserved 68 times on 34 days in 10351 station-times (0.7%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.15
0.16
0.07
0.10
0.18
IFS
0.68
0.62
0.72
0.48
0.63
GFS
0.68
0.60
0.53
0.54
0.52
10 m wind of 10.8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.15
0.14
0.07
0.06
-0.00
IFS
0.31
0.25
0.17
0.06
0.07
GFS
0.34
0.27
0.13
0.06
0.10
10 m wind: bias, in m/sForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
+0.02
+0.02
+0.03
+0.05
−0.02
IFS
+0.01
+0.01
−0.02
+0.03
+0.00
GFS
−0.01
−0.01
+0.01
+0.04
+0.00
10 m wind: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.92
0.89
0.87
0.85
0.88
IFS
0.95
0.94
0.93
0.90
0.84
GFS
1.01
0.98
0.95
0.99
0.93
10 m wind of 8 m/s or more: frequency biasObserved 237 times on 66 days in 8495 station-times (3%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.84
0.79
0.79
0.78
0.82
IFS
0.85
0.92
0.85
0.83
0.67
GFS
1.09
1.05
0.95
1.02
1.05
10 m wind of 8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.45
0.36
0.28
0.21
0.09
IFS
0.47
0.39
0.28
0.18
0.09
GFS
0.42
0.36
0.25
0.17
0.12
10 m wind of 10.8 m/s or more: frequency biasObserved 62 times on 31 days in 8495 station-times (0.7%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.39
0.31
0.21
0.24
0.32
IFS
0.60
0.55
0.66
0.47
0.56
GFS
0.58
0.58
0.61
0.61
0.61
10 m wind of 10.8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.32
0.26
0.14
0.07
0.02
IFS
0.32
0.29
0.18
0.07
0.07
GFS
0.30
0.25
0.15
0.07
0.12
Daily solar energy: bias, in kWh/m²/dayForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
+0.07
+0.05
+0.05
+0.06
+0.09
IFS
+0.25
+0.30
+0.31
–
–
GFS
+0.14
+0.26
+0.26
+0.31
+0.13
Daily solar energy: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.77
0.75
0.73
0.66
0.65
IFS
0.79
0.76
0.76
–
–
GFS
0.99
0.95
0.92
0.85
0.86
Daily solar energy: bias, in kWh/m²/dayForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
+0.01
+0.01
−0.01
−0.04
+0.01
IFS
−0.02
−0.02
−0.05
–
–
GFS
+0.01
−0.01
−0.02
−0.06
−0.04
Daily solar energy: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.76
0.73
0.74
0.68
0.68
IFS
0.77
0.73
0.73
–
–
GFS
0.97
0.93
0.91
0.84
0.86
Against ICON, up to 7 November 2026A closed set that no longer grows
ICON's only archive of past runs was Open-Meteo's. Its last valid day read here is 7 November 2026, and the archive is no longer read. It is compared with AIFS here on the days it has, as a set of its own that no longer grows, while the other sets move forward each week. It returns as a new series from DWD's own files once enough runs have been collected.
2 m temperature: skill against climatology
Raw forecasts, 170 days, 23 May 2026 to 8 November 2026. This sample no longer grows.
2 m temperature: skill against persistence
Raw forecasts, 170 days, 23 May 2026 to 8 November 2026, on the 10,001 of 10,161 station-times where persistence can be scored as well. This sample no longer grows.
2 m temperature: skill against climatology
Bias-corrected forecasts, 140 days, 22 June 2026 to 8 November 2026. This sample no longer grows.
2 m temperature: skill against persistence
Bias-corrected forecasts, 140 days, 22 June 2026 to 8 November 2026, on the 8,216 of 8,362 station-times where persistence can be scored as well. This sample no longer grows.
2 m temperature, AIFS 2: how AIFS's error compared, over 170 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
ICON
▼−12%, AIFS lower error, tested
▼−16%, AIFS lower error, tested
▼−16%, AIFS lower error, tested
–
–
2 m temperature, AIFS 2: how AIFS's error compared, over 140 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
ICON
−15%, not separated
−18%, not separated
−17%, not separated
–
–
10 m wind speed: skill against climatology
Raw forecasts, 170 days, 23 May 2026 to 8 November 2026. This sample no longer grows.
10 m wind speed: skill against persistence
Raw forecasts, 170 days, 23 May 2026 to 8 November 2026, on the 9,848 of 10,118 station-times where persistence can be scored as well. This sample no longer grows.
10 m wind speed: skill against climatology
Bias-corrected forecasts, 140 days, 22 June 2026 to 8 November 2026. This sample no longer grows.
10 m wind speed: skill against persistence
Bias-corrected forecasts, 140 days, 22 June 2026 to 8 November 2026, on the 8,014 of 8,262 station-times where persistence can be scored as well. This sample no longer grows.
10 m wind speed, AIFS 2: how AIFS's error compared, over 170 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
ICON
▼−11%, AIFS lower error, tested
▼−12%, AIFS lower error, tested
▼−17%, AIFS lower error, tested
–
–
10 m wind speed, AIFS 2: how AIFS's error compared, over 140 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
ICON
−4%, not separated
−8%, not separated
−15%, not separated
–
–
Solar energy: skill against climatology
Raw forecasts, 169 days, 23 May 2026 to 7 November 2026. This sample no longer grows.
Solar energy: skill against persistence
Raw forecasts, 169 days, 23 May 2026 to 7 November 2026, on the 4,568 of 4,973 station-times where persistence can be scored as well. This sample no longer grows.
Solar energy: skill against climatology
Bias-corrected forecasts, 138 days, 23 June 2026 to 7 November 2026. This sample no longer grows.
Solar energy: skill against persistence
Bias-corrected forecasts, 137 days, 24 June 2026 to 7 November 2026: those of the 138 days scored on which persistence can be scored as well. This sample no longer grows.
▼AIFS lower error, tested (better)
▲AIFS higher error, tested (worse)
Not separated: no clear difference, too few days, or the verdict depends on how the station is read or on the height correction (hover says which)
ICON's 10 m wind ran low at every lead at these stations, on average by 0.59 to 0.62 m/s. One day ahead, AIFS's error was 11% below ICON's in raw forecasts and 4% below after bias correction, so the raw comparison mostly reflects that offset. The cause of the offset is not established.
2 m temperature: bias, in °CForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
+0.08
+0.09
+0.22
+0.45
−0.14
ICON
+0.11
+0.06
+0.19
–
–
2 m temperature: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.98
0.97
0.99
0.98
0.95
ICON
0.97
0.95
0.97
–
–
2 m temperature of 0 °C or belowObserved 15 times on 5 days in 10161 station-times (0.1%). Too few events to score: at least 30, on at least 10 days, are needed.
2 m temperature: bias, in °CForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
−0.08
−0.07
−0.11
−0.26
−0.13
ICON
−0.00
+0.02
−0.00
–
–
2 m temperature: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.95
0.94
0.96
0.94
0.94
ICON
0.96
0.95
0.98
–
–
2 m temperature of 0 °C or belowObserved 15 times on 5 days in 8362 station-times (0.2%). Too few events to score: at least 30, on at least 10 days, are needed.
10 m wind: bias, in m/sForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
−0.26
−0.34
−0.39
−0.46
−0.42
ICON
−0.62
−0.61
−0.59
–
–
10 m wind: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.92
0.89
0.85
0.83
0.87
ICON
0.87
0.88
0.85
–
–
10 m wind of 8 m/s or more: frequency biasObserved 283 times on 77 days in 10118 station-times (3%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.46
0.38
0.33
0.26
0.37
ICON
0.87
0.93
0.85
–
–
10 m wind of 8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.30
0.23
0.18
0.12
0.05
ICON
0.38
0.31
0.22
–
–
10 m wind of 10.8 m/s or more: frequency biasObserved 66 times on 33 days in 10118 station-times (0.7%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.15
0.17
0.08
0.11
0.18
ICON
1.23
1.33
1.27
–
–
10 m wind of 10.8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.15
0.15
0.07
0.06
-0.00
ICON
0.36
0.30
0.18
–
–
10 m wind: bias, in m/sForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
+0.02
+0.03
+0.03
+0.05
−0.02
ICON
+0.00
0.00
0.00
–
–
10 m wind: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.92
0.90
0.87
0.86
0.88
ICON
0.86
0.88
0.86
–
–
10 m wind of 8 m/s or more: frequency biasObserved 235 times on 65 days in 8262 station-times (3%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.84
0.79
0.78
0.78
0.83
ICON
0.83
0.85
0.81
–
–
10 m wind of 8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.45
0.36
0.28
0.21
0.09
ICON
0.44
0.35
0.24
–
–
10 m wind of 10.8 m/s or more: frequency biasObserved 60 times on 30 days in 8262 station-times (0.7%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.40
0.32
0.22
0.25
0.33
ICON
0.87
0.87
0.95
–
–
10 m wind of 10.8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.33
0.27
0.14
0.07
0.02
ICON
0.36
0.27
0.18
–
–
Daily solar energy: bias, in kWh/m²/dayForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
+0.07
+0.05
+0.05
+0.06
+0.08
ICON
−0.23
−0.23
–
–
–
Daily solar energy: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.77
0.75
0.73
0.66
0.65
ICON
0.91
0.90
–
–
–
Daily solar energy: bias, in kWh/m²/dayForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
+0.01
+0.01
−0.01
−0.04
+0.02
ICON
+0.06
+0.07
–
–
–
Daily solar energy: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.76
0.73
0.74
0.68
0.68
ICON
0.91
0.90
–
–
–
Month by month, three days aheadA pattern to read, without intervals
Monthly values carry no intervals, so read them as a pattern and not as results. November 2026 holds 12 days, fewer than 20.
2 m temperature: raw skill against persistence
10 m wind speed: raw skill against persistence
Solar energy: raw skill against persistence
All scores: 2 m temperature, rawRMSE, MAE, bias and skill for every model and lead
°C, with 95% intervals in brackets
Series
Lead
RMSE
MAE
Bias
Skill against persistence
Skill against climatology
AIFS
Day 1
1.38 (1.28 to 1.46)
1.04 (0.97 to 1.10)
0.08 (-0.07 to 0.23)
61% (57% to 64%)
66% (62% to 70%)
IFS
Day 1
1.52 (1.41 to 1.63)
1.12 (1.04 to 1.21)
0.11 (0.02 to 0.21)
56% (52% to 60%)
63% (58% to 67%)
GFS
Day 1
1.93 (1.79 to 2.07)
1.47 (1.35 to 1.59)
-0.14 (-0.44 to 0.15)
45% (39% to 50%)
53% (46% to 58%)
Persistence
Day 1
3.50 (3.22 to 3.76)
2.73 (2.49 to 2.96)
0.06 (-0.16 to 0.26)
14% (−1% to 27%)
Climatology
Day 1
4.09 (3.46 to 4.74)
3.19 (2.70 to 3.73)
-0.94 (-1.90 to -0.02)
−17% (−36% to 1%)
AIFS
Day 3
1.60 (1.50 to 1.69)
1.20 (1.13 to 1.27)
0.08 (-0.06 to 0.24)
66% (62% to 68%)
61% (55% to 65%)
IFS
Day 3
1.84 (1.72 to 1.96)
1.37 (1.28 to 1.45)
0.14 (-0.01 to 0.31)
61% (56% to 64%)
55% (48% to 61%)
GFS
Day 3
2.24 (2.09 to 2.38)
1.72 (1.60 to 1.84)
-0.08 (-0.38 to 0.19)
52% (46% to 56%)
45% (37% to 52%)
Persistence
Day 3
4.67 (4.10 to 5.20)
3.63 (3.20 to 4.08)
0.17 (-0.41 to 0.71)
−14% (−30% to −1%)
Climatology
Day 3
4.09 (3.45 to 4.74)
3.19 (2.70 to 3.73)
-0.94 (-1.89 to 0.00)
13% (1% to 23%)
AIFS
Day 5
2.13 (1.97 to 2.29)
1.60 (1.48 to 1.73)
0.21 (-0.01 to 0.45)
57% (53% to 61%)
48% (40% to 54%)
IFS
Day 5
2.45 (2.26 to 2.64)
1.83 (1.71 to 1.96)
0.36 (0.10 to 0.63)
51% (45% to 56%)
40% (31% to 48%)
GFS
Day 5
2.80 (2.59 to 3.00)
2.14 (1.99 to 2.30)
0.02 (-0.45 to 0.42)
44% (38% to 49%)
31% (21% to 40%)
Persistence
Day 5
5.02 (4.51 to 5.56)
3.96 (3.56 to 4.39)
0.25 (-0.65 to 1.06)
−23% (−38% to −10%)
Climatology
Day 5
4.09 (3.46 to 4.74)
3.19 (2.71 to 3.73)
-0.94 (-1.89 to -0.01)
19% (9% to 28%)
AIFS
Day 7
2.91 (2.68 to 3.12)
2.20 (2.01 to 2.38)
0.43 (0.09 to 0.82)
48% (41% to 54%)
29% (17% to 38%)
IFS
Day 7
3.20 (2.86 to 3.52)
2.46 (2.20 to 2.71)
0.53 (0.17 to 0.91)
43% (37% to 48%)
22% (12% to 30%)
GFS
Day 7
3.64 (3.20 to 4.06)
2.76 (2.44 to 3.09)
-0.02 (-0.53 to 0.45)
35% (28% to 42%)
11% (−1% to 21%)
Persistence
Day 7
5.60 (4.90 to 6.27)
4.44 (3.88 to 5.01)
0.31 (-0.87 to 1.38)
−37% (−52% to −23%)
Climatology
Day 7
4.09 (3.46 to 4.73)
3.19 (2.70 to 3.73)
-0.94 (-1.89 to -0.01)
27% (19% to 34%)
AIFS
Day 10
3.93 (3.51 to 4.38)
3.03 (2.71 to 3.38)
-0.15 (-0.66 to 0.37)
36% (25% to 44%)
4% (−10% to 14%)
IFS
Day 10
4.44 (3.97 to 4.89)
3.43 (3.09 to 3.77)
0.47 (-0.06 to 1.04)
27% (17% to 36%)
−9% (−22% to 2%)
GFS
Day 10
4.62 (4.06 to 5.14)
3.61 (3.19 to 4.03)
-0.21 (-0.90 to 0.47)
24% (14% to 33%)
−13% (−26% to −3%)
Persistence
Day 10
6.11 (5.17 to 7.06)
4.78 (4.06 to 5.59)
0.34 (-1.21 to 1.72)
−50% (−66% to −35%)
Climatology
Day 10
4.09 (3.46 to 4.73)
3.19 (2.70 to 3.72)
-0.94 (-1.90 to -0.00)
33% (26% to 40%)
Persistence's own row and the skill against it rest on 174 days, 23 May 2026 to 12 November 2026, on the 10,241 of 10,401 station-times where persistence can be scored as well. Every other score needs no earlier measurement.
All scores: 2 m temperature, bias-correctedRMSE, MAE, bias and skill for every model and lead
°C, with 95% intervals in brackets
Series
Lead
RMSE
MAE
Bias
Skill against persistence
Skill against climatology
AIFS
Day 1
1.20 (1.11 to 1.28)
0.90 (0.84 to 0.96)
-0.07 (-0.14 to 0.00)
66% (61% to 69%)
69% (65% to 73%)
IFS
Day 1
1.38 (1.28 to 1.47)
1.03 (0.96 to 1.10)
-0.01 (-0.09 to 0.08)
61% (56% to 64%)
65% (61% to 68%)
GFS
Day 1
1.71 (1.60 to 1.81)
1.29 (1.21 to 1.37)
0.17 (0.02 to 0.34)
51% (44% to 56%)
57% (51% to 61%)
Persistence
Day 1
3.51 (3.14 to 3.85)
2.71 (2.40 to 3.02)
0.06 (-0.17 to 0.28)
11% (−4% to 23%)
Climatology
Day 1
3.93 (3.37 to 4.47)
3.08 (2.66 to 3.53)
0.28 (-0.73 to 1.26)
−12% (−29% to 4%)
AIFS
Day 3
1.49 (1.39 to 1.58)
1.13 (1.06 to 1.19)
-0.07 (-0.15 to 0.00)
68% (64% to 71%)
63% (57% to 67%)
IFS
Day 3
1.76 (1.65 to 1.87)
1.32 (1.23 to 1.40)
0.02 (-0.17 to 0.20)
62% (57% to 66%)
56% (49% to 61%)
GFS
Day 3
2.08 (1.96 to 2.19)
1.57 (1.48 to 1.65)
0.21 (0.01 to 0.40)
56% (49% to 60%)
48% (40% to 54%)
Persistence
Day 3
4.66 (4.00 to 5.29)
3.60 (3.09 to 4.12)
0.25 (-0.32 to 0.84)
−17% (−34% to −1%)
Climatology
Day 3
3.98 (3.42 to 4.54)
3.14 (2.72 to 3.60)
0.31 (-0.74 to 1.34)
15% (1% to 26%)
AIFS
Day 5
2.08 (1.89 to 2.26)
1.55 (1.42 to 1.68)
-0.13 (-0.33 to 0.08)
57% (52% to 62%)
48% (40% to 55%)
IFS
Day 5
2.41 (2.23 to 2.58)
1.81 (1.69 to 1.93)
-0.01 (-0.28 to 0.26)
51% (44% to 56%)
40% (31% to 48%)
GFS
Day 5
2.67 (2.49 to 2.85)
2.03 (1.91 to 2.15)
0.25 (-0.17 to 0.67)
45% (39% to 51%)
34% (24% to 42%)
Persistence
Day 5
4.89 (4.28 to 5.47)
3.82 (3.33 to 4.33)
0.42 (-0.45 to 1.28)
−22% (−39% to −6%)
Climatology
Day 5
4.02 (3.46 to 4.58)
3.19 (2.77 to 3.66)
0.33 (-0.73 to 1.38)
18% (6% to 28%)
AIFS
Day 7
2.87 (2.60 to 3.12)
2.17 (1.96 to 2.38)
-0.27 (-0.69 to 0.10)
47% (36% to 55%)
29% (17% to 39%)
IFS
Day 7
3.07 (2.74 to 3.39)
2.37 (2.12 to 2.62)
-0.07 (-0.49 to 0.34)
43% (34% to 50%)
24% (13% to 33%)
GFS
Day 7
3.59 (3.12 to 4.04)
2.74 (2.41 to 3.09)
0.12 (-0.55 to 0.78)
34% (24% to 42%)
11% (−0.4% to 21.0%)
Persistence
Day 7
5.42 (4.50 to 6.28)
4.22 (3.52 to 4.95)
0.56 (-0.54 to 1.68)
−34% (−55% to −14%)
Climatology
Day 7
4.04 (3.49 to 4.60)
3.22 (2.79 to 3.69)
0.35 (-0.74 to 1.42)
26% (12% to 36%)
AIFS
Day 10
3.84 (3.52 to 4.18)
2.98 (2.73 to 3.25)
-0.14 (-0.83 to 0.54)
33% (22% to 42%)
5% (−8% to 15%)
IFS
Day 10
4.42 (3.85 to 4.96)
3.40 (3.00 to 3.80)
-0.19 (-0.85 to 0.47)
23% (14% to 31%)
−9% (−22% to 2%)
GFS
Day 10
4.60 (3.96 to 5.19)
3.57 (3.10 to 4.07)
0.05 (-0.82 to 0.93)
20% (9% to 29%)
−14% (−27% to −2%)
Persistence
Day 10
5.77 (4.82 to 6.70)
4.50 (3.73 to 5.35)
0.76 (-0.61 to 2.14)
−43% (−66% to −23%)
Climatology
Day 10
4.04 (3.52 to 4.59)
3.23 (2.82 to 3.67)
0.39 (-0.72 to 1.45)
30% (19% to 40%)
Persistence's own row and the skill against it rest on 144 days, 22 June 2026 to 12 November 2026, on the 8,456 of 8,602 station-times where persistence can be scored as well. Every other score needs no earlier measurement.
All scores: 10 m wind speed, rawRMSE, MAE, bias and skill for every model and lead
m/s, with 95% intervals in brackets
Series
Lead
RMSE
MAE
Bias
Skill against persistence
Skill against climatology
AIFS
Day 1
1.20 (1.15 to 1.24)
0.92 (0.88 to 0.95)
-0.26 (-0.32 to -0.19)
39% (34% to 44%)
30% (22% to 38%)
IFS
Day 1
1.21 (1.16 to 1.26)
0.91 (0.88 to 0.95)
-0.21 (-0.27 to -0.16)
39% (34% to 43%)
29% (22% to 36%)
GFS
Day 1
1.25 (1.19 to 1.31)
0.94 (0.89 to 0.98)
0.02 (-0.04 to 0.08)
37% (32% to 40%)
27% (19% to 34%)
Persistence
Day 1
1.98 (1.85 to 2.12)
1.47 (1.36 to 1.58)
-0.00 (-0.07 to 0.07)
−17% (−25% to −9%)
Climatology
Day 1
1.71 (1.55 to 1.90)
1.31 (1.20 to 1.44)
-0.02 (-0.28 to 0.25)
14% (9% to 20%)
AIFS
Day 3
1.31 (1.24 to 1.37)
0.99 (0.94 to 1.03)
-0.34 (-0.42 to -0.27)
42% (36% to 48%)
24% (14% to 32%)
IFS
Day 3
1.35 (1.28 to 1.42)
1.01 (0.96 to 1.06)
-0.26 (-0.32 to -0.20)
40% (35% to 46%)
21% (13% to 29%)
GFS
Day 3
1.41 (1.35 to 1.48)
1.06 (1.01 to 1.11)
-0.06 (-0.11 to 0.00)
37% (32% to 42%)
17% (10% to 24%)
Persistence
Day 3
2.26 (2.07 to 2.48)
1.67 (1.52 to 1.85)
0.01 (-0.15 to 0.17)
−33% (−41% to −26%)
Climatology
Day 3
1.71 (1.55 to 1.90)
1.31 (1.20 to 1.44)
-0.02 (-0.29 to 0.25)
25% (20% to 29%)
AIFS
Day 5
1.50 (1.43 to 1.57)
1.12 (1.07 to 1.17)
-0.38 (-0.47 to -0.30)
36% (28% to 44%)
12% (3% to 21%)
IFS
Day 5
1.60 (1.52 to 1.69)
1.21 (1.15 to 1.27)
-0.27 (-0.37 to -0.17)
32% (24% to 40%)
6% (−3% to 15%)
GFS
Day 5
1.67 (1.57 to 1.76)
1.25 (1.18 to 1.32)
-0.07 (-0.16 to 0.02)
29% (21% to 37%)
3% (−7% to 11%)
Persistence
Day 5
2.36 (2.11 to 2.65)
1.74 (1.57 to 1.96)
0.01 (-0.21 to 0.24)
−39% (−49% to −29%)
Climatology
Day 5
1.71 (1.55 to 1.89)
1.31 (1.20 to 1.44)
-0.02 (-0.28 to 0.26)
28% (23% to 33%)
AIFS
Day 7
1.80 (1.68 to 1.92)
1.33 (1.26 to 1.42)
-0.45 (-0.58 to -0.31)
27% (19% to 34%)
−5% (−13% to 2%)
IFS
Day 7
1.90 (1.79 to 2.01)
1.42 (1.34 to 1.51)
-0.34 (-0.48 to -0.18)
23% (15% to 30%)
−11% (−20% to −3%)
GFS
Day 7
1.99 (1.86 to 2.14)
1.49 (1.39 to 1.61)
-0.06 (-0.21 to 0.11)
19% (12% to 25%)
−16% (−24% to −10%)
Persistence
Day 7
2.45 (2.21 to 2.72)
1.83 (1.66 to 2.04)
0.03 (-0.25 to 0.31)
−44% (−55% to −34%)
Climatology
Day 7
1.71 (1.55 to 1.90)
1.31 (1.20 to 1.44)
-0.02 (-0.28 to 0.25)
31% (26% to 35%)
AIFS
Day 10
2.19 (2.00 to 2.39)
1.63 (1.51 to 1.78)
-0.42 (-0.64 to -0.19)
9% (1% to 16%)
−28% (−36% to −21%)
IFS
Day 10
2.19 (2.02 to 2.38)
1.64 (1.51 to 1.77)
-0.38 (-0.60 to -0.15)
9% (2% to 16%)
−28% (−36% to −21%)
GFS
Day 10
2.26 (2.09 to 2.46)
1.67 (1.56 to 1.81)
-0.08 (-0.28 to 0.12)
6% (−1% to 13%)
−32% (−39% to −25%)
Persistence
Day 10
2.39 (2.19 to 2.60)
1.76 (1.62 to 1.92)
0.02 (-0.29 to 0.33)
−41% (−51% to −32%)
Climatology
Day 10
1.71 (1.55 to 1.90)
1.31 (1.20 to 1.44)
-0.02 (-0.28 to 0.24)
29% (24% to 34%)
Persistence's own row and the skill against it rest on 174 days, 23 May 2026 to 12 November 2026, on the 10,081 of 10,351 station-times where persistence can be scored as well. Every other score needs no earlier measurement.
All scores: 10 m wind speed, bias-correctedRMSE, MAE, bias and skill for every model and lead
m/s, with 95% intervals in brackets
Series
Lead
RMSE
MAE
Bias
Skill against persistence
Skill against climatology
AIFS
Day 1
1.02 (0.97 to 1.07)
0.77 (0.73 to 0.81)
0.02 (-0.04 to 0.07)
49% (45% to 52%)
43% (36% to 50%)
IFS
Day 1
1.03 (0.98 to 1.09)
0.78 (0.74 to 0.82)
0.01 (-0.06 to 0.06)
48% (44% to 52%)
43% (35% to 49%)
GFS
Day 1
1.13 (1.07 to 1.19)
0.84 (0.80 to 0.89)
-0.01 (-0.08 to 0.07)
43% (40% to 47%)
37% (30% to 44%)
Persistence
Day 1
2.00 (1.84 to 2.16)
1.47 (1.34 to 1.61)
-0.00 (-0.09 to 0.08)
−12% (−23% to −3%)
Climatology
Day 1
1.80 (1.59 to 2.04)
1.38 (1.23 to 1.54)
0.05 (-0.29 to 0.41)
11% (3% to 18%)
AIFS
Day 3
1.14 (1.07 to 1.21)
0.86 (0.81 to 0.91)
0.02 (-0.04 to 0.09)
51% (45% to 56%)
38% (29% to 45%)
IFS
Day 3
1.20 (1.13 to 1.28)
0.90 (0.84 to 0.95)
0.01 (-0.06 to 0.08)
48% (43% to 53%)
34% (25% to 41%)
GFS
Day 3
1.32 (1.24 to 1.40)
0.98 (0.93 to 1.04)
-0.01 (-0.10 to 0.07)
43% (38% to 48%)
28% (19% to 35%)
Persistence
Day 3
2.31 (2.08 to 2.59)
1.71 (1.53 to 1.93)
0.01 (-0.19 to 0.22)
−28% (−38% to −19%)
Climatology
Day 3
1.82 (1.60 to 2.07)
1.39 (1.24 to 1.56)
0.05 (-0.31 to 0.42)
22% (16% to 28%)
AIFS
Day 5
1.37 (1.30 to 1.44)
1.03 (0.98 to 1.08)
0.03 (-0.07 to 0.13)
44% (35% to 51%)
25% (15% to 33%)
IFS
Day 5
1.50 (1.41 to 1.60)
1.13 (1.06 to 1.20)
-0.02 (-0.14 to 0.10)
38% (30% to 46%)
17% (7% to 26%)
GFS
Day 5
1.58 (1.48 to 1.68)
1.18 (1.11 to 1.26)
0.01 (-0.09 to 0.10)
35% (25% to 43%)
13% (1% to 22%)
Persistence
Day 5
2.43 (2.12 to 2.79)
1.79 (1.57 to 2.05)
0.03 (-0.27 to 0.33)
−35% (−48% to −24%)
Climatology
Day 5
1.82 (1.60 to 2.06)
1.39 (1.23 to 1.55)
0.05 (-0.31 to 0.42)
26% (19% to 32%)
AIFS
Day 7
1.68 (1.53 to 1.85)
1.26 (1.17 to 1.37)
0.05 (-0.09 to 0.19)
34% (27% to 40%)
7% (−1% to 14%)
IFS
Day 7
1.81 (1.68 to 1.93)
1.35 (1.26 to 1.45)
0.03 (-0.14 to 0.22)
29% (21% to 36%)
0% (−10% to 9%)
GFS
Day 7
1.97 (1.80 to 2.14)
1.47 (1.35 to 1.61)
0.04 (-0.16 to 0.24)
22% (14% to 29%)
−9% (−18.6% to −0.5%)
Persistence
Day 7
2.51 (2.23 to 2.82)
1.87 (1.66 to 2.11)
0.05 (-0.32 to 0.41)
−41% (−55% to −28%)
Climatology
Day 7
1.81 (1.59 to 2.04)
1.38 (1.23 to 1.54)
0.05 (-0.32 to 0.43)
29% (22% to 35%)
AIFS
Day 10
2.09 (1.86 to 2.35)
1.57 (1.42 to 1.76)
-0.02 (-0.32 to 0.26)
16% (8% to 25%)
−16% (−24% to −8%)
IFS
Day 10
2.12 (1.94 to 2.33)
1.59 (1.47 to 1.73)
0.00 (-0.34 to 0.35)
15% (8% to 23%)
−17% (−26% to −10%)
GFS
Day 10
2.21 (2.00 to 2.47)
1.64 (1.50 to 1.82)
0.00 (-0.29 to 0.27)
11% (4% to 19%)
−23% (−31% to −14%)
Persistence
Day 10
2.48 (2.25 to 2.73)
1.82 (1.65 to 2.01)
0.08 (-0.35 to 0.52)
−39% (−52% to −28%)
Climatology
Day 10
1.80 (1.59 to 2.04)
1.37 (1.22 to 1.54)
0.04 (-0.33 to 0.43)
28% (22% to 34%)
Persistence's own row and the skill against it rest on 144 days, 22 June 2026 to 12 November 2026, on the 8,247 of 8,495 station-times where persistence can be scored as well. Every other score needs no earlier measurement.
All scores: solar energy, rawRMSE, MAE, bias and skill for every model and lead
kWh/m²/day, with 95% intervals in brackets
Series
Lead
RMSE
MAE
Bias
Skill against persistence
Skill against climatology
AIFS
Day 1
0.62 (0.52 to 0.70)
0.47 (0.40 to 0.54)
0.07 (0.01 to 0.13)
66% (63% to 68%)
58% (54% to 62%)
IFS
Day 1
0.73 (0.60 to 0.84)
0.52 (0.43 to 0.61)
0.25 (0.17 to 0.34)
60% (57% to 62%)
51% (45% to 56%)
GFS
Day 1
0.81 (0.68 to 0.93)
0.57 (0.49 to 0.66)
0.14 (0.06 to 0.23)
55% (50% to 59%)
46% (39% to 52%)
Persistence
Day 1
1.79 (1.50 to 2.05)
1.34 (1.12 to 1.56)
0.06 (-0.07 to 0.19)
−21% (−33% to −9%)
Climatology
Day 1
1.49 (1.23 to 1.71)
1.22 (0.99 to 1.44)
-0.33 (-0.60 to -0.06)
17% (8% to 25%)
AIFS
Day 3
0.78 (0.65 to 0.89)
0.58 (0.49 to 0.67)
0.05 (-0.02 to 0.13)
60% (56% to 63%)
47% (41% to 53%)
IFS
Day 3
0.94 (0.78 to 1.06)
0.67 (0.55 to 0.77)
0.30 (0.20 to 0.41)
52% (46% to 56%)
37% (29% to 45%)
GFS
Day 3
1.10 (0.90 to 1.29)
0.77 (0.64 to 0.91)
0.26 (0.13 to 0.38)
43% (38% to 48%)
26% (17% to 35%)
Persistence
Day 3
1.93 (1.58 to 2.23)
1.47 (1.21 to 1.74)
0.11 (-0.12 to 0.32)
−29% (−41% to −18%)
Climatology
Day 3
1.49 (1.23 to 1.71)
1.22 (0.98 to 1.44)
-0.33 (-0.60 to -0.06)
23% (15% to 29%)
AIFS
Day 5
1.01 (0.83 to 1.16)
0.75 (0.62 to 0.87)
0.05 (-0.04 to 0.15)
52% (48% to 56%)
32% (25% to 40%)
IFS
Day 5
1.16 (0.96 to 1.34)
0.84 (0.70 to 0.97)
0.31 (0.17 to 0.46)
44% (39% to 50%)
22% (11% to 32%)
GFS
Day 5
1.34 (1.10 to 1.55)
0.96 (0.80 to 1.14)
0.26 (0.13 to 0.39)
36% (31% to 42%)
10% (−1% to 20%)
Persistence
Day 5
2.08 (1.73 to 2.38)
1.60 (1.32 to 1.87)
0.16 (-0.14 to 0.45)
−40% (−52% to −29%)
Climatology
Day 5
1.49 (1.23 to 1.70)
1.22 (0.98 to 1.44)
-0.33 (-0.60 to -0.06)
28% (23% to 34%)
AIFS
Day 7
1.26 (1.04 to 1.45)
0.94 (0.78 to 1.10)
0.06 (-0.10 to 0.22)
42% (34% to 49%)
16% (6% to 26%)
GFS
Day 7
1.55 (1.25 to 1.82)
1.11 (0.91 to 1.32)
0.31 (0.11 to 0.53)
28% (19% to 36%)
−4% (−16% to 8%)
Persistence
Day 7
2.14 (1.75 to 2.48)
1.65 (1.33 to 1.96)
0.20 (-0.18 to 0.55)
−44% (−56% to −33%)
Climatology
Day 7
1.49 (1.23 to 1.71)
1.22 (0.99 to 1.44)
-0.33 (-0.60 to -0.06)
31% (25% to 36%)
AIFS
Day 10
1.45 (1.19 to 1.68)
1.09 (0.90 to 1.28)
0.09 (-0.05 to 0.23)
35% (27% to 43%)
3% (−7% to 13%)
GFS
Day 10
1.67 (1.38 to 1.93)
1.26 (1.05 to 1.48)
0.13 (-0.07 to 0.34)
25% (17% to 32%)
−12% (−23% to −2%)
Persistence
Day 10
2.22 (1.83 to 2.58)
1.73 (1.41 to 2.05)
0.24 (-0.21 to 0.66)
−50% (−60% to −38%)
Climatology
Day 10
1.49 (1.23 to 1.71)
1.22 (0.98 to 1.44)
-0.33 (-0.60 to -0.06)
33% (28% to 37%)
Persistence's own row and the skill against it rest on 174 days, 23 May 2026 to 12 November 2026, on the 4,708 of 5,123 station-times where persistence can be scored as well. Every other score needs no earlier measurement.
All scores: solar energy, bias-correctedRMSE, MAE, bias and skill for every model and lead
kWh/m²/day, with 95% intervals in brackets
Series
Lead
RMSE
MAE
Bias
Skill against persistence
Skill against climatology
AIFS
Day 1
0.59 (0.49 to 0.68)
0.45 (0.38 to 0.52)
0.01 (-0.07 to 0.08)
67% (64% to 69%)
57% (53% to 61%)
IFS
Day 1
0.66 (0.54 to 0.77)
0.51 (0.42 to 0.59)
-0.02 (-0.10 to 0.06)
63% (60% to 66%)
52% (47% to 56%)
GFS
Day 1
0.76 (0.64 to 0.87)
0.56 (0.47 to 0.64)
0.01 (-0.07 to 0.10)
58% (52% to 62%)
45% (37% to 51%)
Persistence
Day 1
1.79 (1.43 to 2.10)
1.34 (1.08 to 1.61)
0.03 (-0.10 to 0.17)
−31% (−41% to −21%)
Climatology
Day 1
1.37 (1.10 to 1.61)
1.09 (0.87 to 1.31)
0.01 (-0.25 to 0.29)
24% (17% to 29%)
AIFS
Day 3
0.78 (0.63 to 0.91)
0.57 (0.47 to 0.67)
0.01 (-0.08 to 0.11)
59% (54% to 63%)
44% (40% to 47%)
IFS
Day 3
0.88 (0.73 to 1.02)
0.67 (0.55 to 0.78)
-0.02 (-0.13 to 0.08)
53% (46% to 57%)
36% (29% to 42%)
GFS
Day 3
1.02 (0.85 to 1.20)
0.75 (0.64 to 0.87)
-0.01 (-0.15 to 0.14)
45% (38% to 50%)
26% (17% to 34%)
Persistence
Day 3
1.85 (1.46 to 2.23)
1.40 (1.12 to 1.73)
0.06 (-0.15 to 0.29)
−35% (−46% to −24%)
Climatology
Day 3
1.38 (1.11 to 1.62)
1.10 (0.89 to 1.31)
0.01 (-0.25 to 0.31)
26% (19% to 32%)
AIFS
Day 5
1.00 (0.82 to 1.17)
0.74 (0.61 to 0.87)
-0.01 (-0.14 to 0.14)
49% (44% to 53%)
27% (21% to 32%)
IFS
Day 5
1.12 (0.93 to 1.30)
0.85 (0.71 to 0.98)
-0.05 (-0.22 to 0.11)
42% (37% to 47%)
19% (12% to 24%)
GFS
Day 5
1.28 (1.04 to 1.50)
0.93 (0.78 to 1.10)
-0.02 (-0.16 to 0.12)
35% (29% to 39%)
8% (1% to 14%)
Persistence
Day 5
1.95 (1.53 to 2.33)
1.49 (1.17 to 1.82)
0.10 (-0.18 to 0.40)
−42% (−52% to −32%)
Climatology
Day 5
1.38 (1.11 to 1.62)
1.10 (0.89 to 1.31)
0.02 (-0.27 to 0.32)
30% (24% to 34%)
AIFS
Day 7
1.22 (1.03 to 1.41)
0.94 (0.78 to 1.10)
-0.04 (-0.21 to 0.13)
39% (31% to 44%)
12% (4% to 19%)
GFS
Day 7
1.45 (1.17 to 1.70)
1.08 (0.87 to 1.29)
-0.06 (-0.27 to 0.16)
27% (20% to 34%)
−4% (−15% to 7%)
Persistence
Day 7
1.99 (1.55 to 2.37)
1.50 (1.17 to 1.84)
0.10 (-0.23 to 0.46)
−43% (−53% to −33%)
Climatology
Day 7
1.39 (1.13 to 1.63)
1.11 (0.90 to 1.32)
0.02 (-0.27 to 0.34)
30% (25% to 35%)
AIFS
Day 10
1.39 (1.12 to 1.64)
1.05 (0.85 to 1.26)
0.01 (-0.14 to 0.16)
33% (28% to 39%)
1% (−9% to 10%)
GFS
Day 10
1.54 (1.26 to 1.79)
1.19 (0.97 to 1.41)
-0.04 (-0.23 to 0.15)
26% (19% to 33%)
−10% (−20% to −1%)
Persistence
Day 10
2.09 (1.67 to 2.48)
1.59 (1.27 to 1.93)
0.14 (-0.24 to 0.57)
−49% (−60% to −38%)
Climatology
Day 10
1.41 (1.14 to 1.65)
1.12 (0.91 to 1.33)
0.02 (-0.29 to 0.36)
33% (28% to 38%)
Persistence's own row and the skill against it rest on 142 days, 24 June 2026 to 12 November 2026: those of the 143 days scored on which persistence can be scored as well. Every other score needs no earlier measurement.
AIFS 1.1: in service from 27 August 2025; forecasts issued from 28 August 2025 to 11 May 2026.
2 m temperature: skill against climatology
Raw forecasts, 248 days, 7 September 2025 to 12 May 2026. Hover for values; AIFS shows its 95% range, others on hover.
2 m temperature: skill against persistence
Raw forecasts, 248 days, 7 September 2025 to 12 May 2026, on the 14,636 of 14,824 station-times where persistence can be scored as well. Hover for values; AIFS shows its 95% range, others on hover.
2 m temperature: skill against climatology
Bias-corrected forecasts, 218 days, 7 October 2025 to 12 May 2026. Hover for values; AIFS shows its 95% range, others on hover.
2 m temperature: skill against persistence
Bias-corrected forecasts, 218 days, 7 October 2025 to 12 May 2026, on the 12,838 of 13,028 station-times where persistence can be scored as well. Hover for values; AIFS shows its 95% range, others on hover.
AIFS's RMSE against each model, in %, over 248 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
▼−18%, AIFS lower error, tested
▼−15%, AIFS lower error, tested
▼−12%, AIFS lower error, tested
−12%, not separated
−9%, not separated
GFS
▼−32%, AIFS lower error, tested
▼−29%, AIFS lower error, tested
▼−23%, AIFS lower error, tested
▼−14%, AIFS lower error, tested
−8%, not separated
AIFS's RMSE against each model, in %, over 218 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
▼−17%, AIFS lower error, tested
▼−16%, AIFS lower error, tested
−11%, not separated
−10%, not separated
−8%, not separated
GFS
▼−34%, AIFS lower error, tested
▼−30%, AIFS lower error, tested
▼−23%, AIFS lower error, tested
−12%, not separated
−5%, not separated
10 m wind speed: skill against climatology
Raw forecasts, 248 days, 7 September 2025 to 12 May 2026. Hover for values; AIFS shows its 95% range, others on hover.
10 m wind speed: skill against persistence
Raw forecasts, 248 days, 7 September 2025 to 12 May 2026, on the 14,623 of 14,824 station-times where persistence can be scored as well. Hover for values; AIFS shows its 95% range, others on hover.
10 m wind speed: skill against climatology
Bias-corrected forecasts, 218 days, 7 October 2025 to 12 May 2026. Hover for values; AIFS shows its 95% range, others on hover.
10 m wind speed: skill against persistence
Bias-corrected forecasts, 218 days, 7 October 2025 to 12 May 2026, on the 12,813 of 13,028 station-times where persistence can be scored as well. Hover for values; AIFS shows its 95% range, others on hover.
AIFS's RMSE against each model, in %, over 248 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
0%, not separated
−2%, not separated
−4%, not separated
−2%, not separated
−8%, not separated
GFS
−4%, not separated
▼−8%, AIFS lower error, tested
▼−10%, AIFS lower error, tested
▼−9%, AIFS lower error, tested
▼−16%, AIFS lower error, tested
AIFS's RMSE against each model, in %, over 218 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
0%, not separated
−4%, not separated
▼−6%, AIFS lower error, tested
−3%, not separated
−9%, not separated
GFS
▼−9%, AIFS lower error, tested
▼−14%, AIFS lower error, tested
▼−15%, AIFS lower error, tested
▼−11%, AIFS lower error, tested
▼−19%, AIFS lower error, tested
Solar energy: skill against climatology
Raw forecasts, 247 days, 7 September 2025 to 12 May 2026. Hover for values; AIFS shows its 95% range, others on hover.
Solar energy: skill against persistence
Raw forecasts, 230 days, 7 September 2025 to 12 May 2026: those of the 247 days scored on which persistence can be scored as well. Hover for values; AIFS shows its 95% range, others on hover.
Solar energy: skill against climatology
Bias-corrected forecasts, 216 days, 8 October 2025 to 12 May 2026. Hover for values; AIFS shows its 95% range, others on hover.
Solar energy: skill against persistence
Bias-corrected forecasts, 197 days, 9 October 2025 to 12 May 2026: those of the 216 days scored on which persistence can be scored as well. Hover for values; AIFS shows its 95% range, others on hover.
AIFS's RMSE against each model, in %, over 247 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
▼−10%, AIFS lower error, tested
−10%, not separated
−10%, not separated
–
–
GFS
▼−26%, AIFS lower error, tested
▼−26%, AIFS lower error, tested
▼−25%, AIFS lower error, tested
▼−19%, AIFS lower error, tested
▼−17%, AIFS lower error, tested
ICON is not compared on solar: its forecasts come from Open-Meteo, which assembles the day differently.
AIFS's RMSE against each model, in %, over 216 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
−4%, not separated
−7%, not separated
−7%, not separated
–
–
GFS
▼−25%, AIFS lower error, tested
▼−24%, AIFS lower error, tested
▼−24%, AIFS lower error, tested
−20%, not separated
▼−16%, AIFS lower error, tested
ICON is not compared on solar: its forecasts come from Open-Meteo, which assembles the day differently.
▼AIFS lower error, tested (better)
▲AIFS higher error, tested (worse)
Not separated: no clear difference, too few days, or the verdict depends on how the station is read or on the height correction (hover says which)
Each number is AIFS's RMSE minus the other model's, as a percentage of the other model's RMSE: minus means AIFS's was lower (better). Hover or tap a cell for the difference in units, its 95% interval and the p-value.
Skill against climatology at day 7, 2 m temperature: AIFS 24%, IFS 13%, GFS 12%. Over these days, AIFS had the highest skill against climatology at this lead.
Skill against climatology at day 7, 10 m wind: AIFS −6%, IFS −8%, GFS −16%. Over these days, no model had a lower RMSE than climatology at this lead, and AIFS came closest to it.
Skill against climatology at day 7, daily solar energy: AIFS 10%, GFS −11%. Over these days, no model had a lower RMSE than climatology at this lead, and AIFS came closest to it.
Bias, activity and eventsWhether a model runs high or low, how much it varies, and how it does in strong wind and frost
What an average error hides: whether a model runs high or low, whether it varies as much as the weather, and how it does when it matters. Bold and underlined marks the value nearest the ideal at each lead, as printed. It is not a verdict: these values are not tested against each other. Hover an event score for its 95% interval. The thresholds were fixed before scoring.
2 m temperature: bias, in °CForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
+0.02
+0.04
+0.01
+0.13
−0.11
IFS
−0.07
−0.07
−0.19
−0.24
−0.46
GFS
+0.15
+0.06
+0.01
+0.10
+0.06
2 m temperature: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.97
0.95
0.96
0.95
1.00
IFS
0.92
0.92
0.93
0.96
0.96
GFS
0.94
0.92
0.94
0.92
0.98
2 m temperature of 0 °C or below: frequency biasObserved 2022 times on 122 days in 14824 station-times (14%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.97
0.99
1.03
0.96
1.02
IFS
0.97
0.98
1.05
1.02
1.06
GFS
0.84
0.84
0.86
0.84
0.85
2 m temperature of 0 °C or below: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.73
0.69
0.60
0.41
0.34
IFS
0.65
0.61
0.53
0.38
0.29
GFS
0.61
0.56
0.48
0.35
0.27
2 m temperature: bias, in °CForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
+0.02
+0.02
+0.01
−0.01
−0.03
IFS
0.00
−0.01
−0.04
−0.10
−0.21
GFS
−0.06
−0.09
−0.12
−0.11
−0.30
2 m temperature: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.96
0.94
0.94
0.93
0.97
IFS
0.93
0.93
0.94
0.97
0.94
GFS
0.96
0.94
0.95
0.93
0.99
2 m temperature of 0 °C or below: frequency biasObserved 2021 times on 121 days in 13028 station-times (16%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.96
0.96
0.99
1.01
1.05
IFS
0.94
0.95
0.95
0.87
1.04
GFS
1.00
0.99
0.98
0.96
1.02
2 m temperature of 0 °C or below: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.72
0.67
0.59
0.39
0.36
IFS
0.66
0.59
0.50
0.35
0.29
GFS
0.60
0.54
0.47
0.34
0.30
10 m wind: bias, in m/sForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
−0.30
−0.38
−0.48
−0.47
−0.42
IFS
−0.19
−0.23
−0.26
−0.23
−0.18
GFS
−0.00
−0.07
−0.07
+0.01
+0.28
10 m wind: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.84
0.84
0.81
0.83
0.85
IFS
0.91
0.91
0.90
0.90
0.95
GFS
1.00
0.98
0.99
1.00
1.13
10 m wind of 8 m/s or more: frequency biasObserved 660 times on 144 days in 14824 station-times (4%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.39
0.33
0.28
0.28
0.34
IFS
0.60
0.58
0.56
0.58
0.65
GFS
1.04
0.89
0.99
1.02
1.53
10 m wind of 8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.29
0.21
0.13
0.08
0.03
IFS
0.29
0.24
0.19
0.12
0.05
GFS
0.37
0.30
0.21
0.10
0.06
10 m wind of 10.8 m/s or more: frequency biasObserved 94 times on 42 days in 14824 station-times (0.6%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.23
0.21
0.15
0.12
0.24
IFS
0.82
0.74
0.76
0.92
0.84
GFS
0.92
0.84
0.89
0.95
2.25
10 m wind of 10.8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.17
0.14
0.07
0.02
-0.00
IFS
0.23
0.20
0.14
0.09
0.04
GFS
0.33
0.20
0.18
0.06
0.04
10 m wind: bias, in m/sForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
−0.02
−0.03
−0.04
−0.01
+0.01
IFS
−0.02
−0.01
+0.01
+0.04
+0.01
GFS
−0.02
−0.01
−0.02
+0.05
+0.05
10 m wind: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.82
0.81
0.77
0.83
0.84
IFS
0.89
0.89
0.88
0.91
0.95
GFS
0.99
0.98
0.98
1.02
1.12
10 m wind of 8 m/s or more: frequency biasObserved 554 times on 128 days in 13028 station-times (4%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.58
0.53
0.44
0.58
0.60
IFS
0.66
0.70
0.64
0.82
0.86
GFS
0.98
0.89
0.95
1.09
1.34
10 m wind of 8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.37
0.28
0.17
0.10
0.05
IFS
0.38
0.31
0.21
0.13
0.05
GFS
0.41
0.31
0.18
0.11
0.04
10 m wind of 10.8 m/s or more: frequency biasObserved 75 times on 35 days in 13028 station-times (0.6%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.28
0.28
0.15
0.35
0.29
IFS
0.45
0.52
0.56
0.75
0.97
GFS
0.75
0.77
0.68
1.07
2.11
10 m wind of 10.8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.23
0.15
0.05
0.02
0.02
IFS
0.27
0.20
0.17
0.07
0.02
GFS
0.35
0.18
0.15
0.10
0.01
Daily solar energy: bias, in kWh/m²/dayForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
+0.02
+0.03
+0.01
+0.01
−0.01
IFS
+0.16
+0.17
+0.15
–
–
GFS
+0.08
+0.10
+0.12
+0.05
−0.05
Daily solar energy: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.82
0.79
0.78
0.71
0.61
IFS
0.84
0.84
0.81
–
–
GFS
1.04
1.01
0.96
0.91
0.85
Daily solar energy: bias, in kWh/m²/dayForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
−0.01
−0.00
−0.01
−0.05
−0.09
IFS
−0.01
−0.01
−0.02
–
–
GFS
+0.01
−0.02
−0.04
−0.04
−0.10
Daily solar energy: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.82
0.79
0.78
0.70
0.63
IFS
0.85
0.84
0.81
–
–
GFS
1.04
1.02
0.94
0.91
0.85
Against ICON, up to 7 November 2026A closed set that no longer grows
ICON's only archive of past runs was Open-Meteo's. Its last valid day read here is 7 November 2026, and the archive is no longer read. It is compared with AIFS here on the days it has, as a set of its own that no longer grows, while the other sets move forward each week. It returns as a new series from DWD's own files once enough runs have been collected.
2 m temperature: skill against climatology
Raw forecasts, 240 days, 7 September 2025 to 12 May 2026. This sample no longer grows.
2 m temperature: skill against persistence
Raw forecasts, 240 days, 7 September 2025 to 12 May 2026, on the 14,162 of 14,344 station-times where persistence can be scored as well. This sample no longer grows.
2 m temperature: skill against climatology
Bias-corrected forecasts, 210 days, 7 October 2025 to 12 May 2026. This sample no longer grows.
2 m temperature: skill against persistence
Bias-corrected forecasts, 210 days, 7 October 2025 to 12 May 2026, on the 12,372 of 12,548 station-times where persistence can be scored as well. This sample no longer grows.
2 m temperature, AIFS 1.1: how AIFS's error compared, over 240 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
ICON
▼−9%, AIFS lower error, tested
▼−13%, AIFS lower error, tested
▼−13%, AIFS lower error, tested
–
–
2 m temperature, AIFS 1.1: how AIFS's error compared, over 210 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
ICON
▼−9%, AIFS lower error, tested
▼−14%, AIFS lower error, tested
▼−14%, AIFS lower error, tested
–
–
10 m wind speed: skill against climatology
Raw forecasts, 248 days, 7 September 2025 to 12 May 2026. This sample no longer grows.
10 m wind speed: skill against persistence
Raw forecasts, 248 days, 7 September 2025 to 12 May 2026, on the 14,623 of 14,824 station-times where persistence can be scored as well. This sample no longer grows.
10 m wind speed: skill against climatology
Bias-corrected forecasts, 218 days, 7 October 2025 to 12 May 2026. This sample no longer grows.
10 m wind speed: skill against persistence
Bias-corrected forecasts, 218 days, 7 October 2025 to 12 May 2026, on the 12,813 of 13,028 station-times where persistence can be scored as well. This sample no longer grows.
10 m wind speed, AIFS 1.1: how AIFS's error compared, over 248 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
ICON
▼−16%, AIFS lower error, tested
▼−14%, AIFS lower error, tested
▼−14%, AIFS lower error, tested
–
–
10 m wind speed, AIFS 1.1: how AIFS's error compared, over 218 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
ICON
▼−6%, AIFS lower error, tested
▼−8%, AIFS lower error, tested
▼−10%, AIFS lower error, tested
–
–
Solar energy: skill against climatology
Raw forecasts, 243 days, 7 September 2025 to 12 May 2026. This sample no longer grows.
Solar energy: skill against persistence
Raw forecasts, 226 days, 7 September 2025 to 12 May 2026: those of the 243 days scored on which persistence can be scored as well. This sample no longer grows.
Solar energy: skill against climatology
Bias-corrected forecasts, 212 days, 8 October 2025 to 12 May 2026. This sample no longer grows.
Solar energy: skill against persistence
Bias-corrected forecasts, 193 days, 9 October 2025 to 12 May 2026: those of the 212 days scored on which persistence can be scored as well. This sample no longer grows.
▼AIFS lower error, tested (better)
▲AIFS higher error, tested (worse)
Not separated: no clear difference, too few days, or the verdict depends on how the station is read or on the height correction (hover says which)
ICON's 10 m wind ran low at every lead at these stations, on average by 0.58 to 0.62 m/s. One day ahead, AIFS's error was 16% below ICON's in raw forecasts and 6% below after bias correction, so the raw comparison mostly reflects that offset. The cause of the offset is not established.
2 m temperature: bias, in °CForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
+0.01
+0.03
−0.01
+0.06
−0.18
ICON
+0.05
+0.09
+0.06
–
–
2 m temperature: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.96
0.95
0.95
0.94
0.97
ICON
0.93
0.92
0.92
–
–
2 m temperature of 0 °C or below: frequency biasObserved 2017 times on 118 days in 14344 station-times (14%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.97
0.99
1.03
0.97
1.02
ICON
0.94
0.93
0.94
–
–
2 m temperature of 0 °C or below: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.73
0.69
0.60
0.41
0.34
ICON
0.72
0.63
0.53
–
–
2 m temperature: bias, in °CForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
+0.01
+0.00
−0.02
−0.10
−0.12
ICON
−0.00
−0.02
−0.04
–
–
2 m temperature: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.96
0.94
0.94
0.92
0.96
ICON
0.93
0.92
0.91
–
–
2 m temperature of 0 °C or below: frequency biasObserved 2016 times on 117 days in 12548 station-times (16%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.96
0.97
0.99
1.01
1.05
ICON
0.97
0.98
1.00
–
–
2 m temperature of 0 °C or below: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.72
0.67
0.59
0.39
0.35
ICON
0.72
0.63
0.51
–
–
10 m wind: bias, in m/sForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
−0.30
−0.38
−0.48
−0.47
−0.42
ICON
−0.58
−0.61
−0.62
–
–
10 m wind: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.84
0.84
0.81
0.83
0.85
ICON
0.88
0.88
0.89
–
–
10 m wind of 8 m/s or more: frequency biasObserved 660 times on 144 days in 14824 station-times (4%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.39
0.33
0.28
0.28
0.34
ICON
0.80
0.75
0.80
–
–
10 m wind of 8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.29
0.21
0.13
0.08
0.03
ICON
0.25
0.21
0.17
–
–
10 m wind of 10.8 m/s or more: frequency biasObserved 94 times on 42 days in 14824 station-times (0.6%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.23
0.21
0.15
0.12
0.24
ICON
1.71
1.61
1.71
–
–
10 m wind of 10.8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.17
0.14
0.07
0.02
-0.00
ICON
0.25
0.17
0.17
–
–
10 m wind: bias, in m/sForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
−0.02
−0.03
−0.04
−0.01
+0.01
ICON
−0.00
+0.01
+0.01
–
–
10 m wind: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.82
0.81
0.77
0.83
0.84
ICON
0.87
0.87
0.87
–
–
10 m wind of 8 m/s or more: frequency biasObserved 554 times on 128 days in 13028 station-times (4%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.58
0.53
0.44
0.58
0.60
ICON
0.71
0.71
0.70
–
–
10 m wind of 8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.37
0.28
0.17
0.10
0.05
ICON
0.34
0.28
0.18
–
–
10 m wind of 10.8 m/s or more: frequency biasObserved 75 times on 35 days in 13028 station-times (0.6%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.28
0.28
0.15
0.35
0.29
ICON
0.85
0.69
0.87
–
–
10 m wind of 10.8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.23
0.15
0.05
0.02
0.02
ICON
0.30
0.21
0.14
–
–
Daily solar energy: bias, in kWh/m²/dayForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
+0.02
+0.02
+0.00
−0.01
−0.02
ICON
−0.04
−0.03
–
–
–
Daily solar energy: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.82
0.79
0.78
0.70
0.61
ICON
0.86
0.86
–
–
–
Daily solar energy: bias, in kWh/m²/dayForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
−0.01
−0.01
−0.02
−0.07
−0.10
ICON
−0.01
−0.00
–
–
–
Daily solar energy: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.82
0.79
0.77
0.70
0.64
ICON
0.84
0.84
–
–
–
Month by month, three days aheadA pattern to read, without intervals
Monthly values carry no intervals, so read them as a pattern and not as results. May 2026 holds 12 days, fewer than 20.
2 m temperature: raw skill against persistence
10 m wind speed: raw skill against persistence
Solar energy: raw skill against persistence
All scores: 2 m temperature, rawRMSE, MAE, bias and skill for every model and lead
°C, with 95% intervals in brackets
Series
Lead
RMSE
MAE
Bias
Skill against persistence
Skill against climatology
AIFS
Day 1
1.23 (1.15 to 1.31)
0.91 (0.85 to 0.97)
0.02 (-0.07 to 0.12)
60% (57% to 63%)
69% (65% to 72%)
IFS
Day 1
1.50 (1.41 to 1.59)
1.11 (1.04 to 1.18)
-0.07 (-0.20 to 0.05)
52% (48% to 56%)
62% (58% to 65%)
GFS
Day 1
1.81 (1.69 to 1.95)
1.34 (1.26 to 1.43)
0.15 (-0.10 to 0.40)
42% (35% to 47%)
54% (48% to 59%)
Persistence
Day 1
3.11 (2.88 to 3.33)
2.37 (2.19 to 2.55)
0.04 (-0.14 to 0.20)
21% (11% to 30%)
Climatology
Day 1
3.94 (3.48 to 4.38)
3.10 (2.72 to 3.51)
0.32 (-0.59 to 1.24)
−27% (−42% to −12%)
AIFS
Day 3
1.52 (1.43 to 1.62)
1.15 (1.07 to 1.22)
0.04 (-0.09 to 0.19)
67% (64% to 69%)
61% (56% to 65%)
IFS
Day 3
1.80 (1.69 to 1.91)
1.35 (1.27 to 1.43)
-0.07 (-0.23 to 0.09)
60% (57% to 64%)
54% (49% to 58%)
GFS
Day 3
2.14 (2.02 to 2.26)
1.62 (1.53 to 1.71)
0.06 (-0.24 to 0.35)
53% (49% to 57%)
46% (39% to 51%)
Persistence
Day 3
4.55 (4.18 to 4.90)
3.55 (3.26 to 3.85)
0.08 (-0.36 to 0.53)
−16% (−30% to −3%)
Climatology
Day 3
3.94 (3.48 to 4.37)
3.10 (2.72 to 3.51)
0.32 (-0.58 to 1.22)
13% (3% to 23%)
AIFS
Day 5
2.12 (1.97 to 2.25)
1.60 (1.48 to 1.72)
0.01 (-0.20 to 0.23)
57% (53% to 60%)
46% (40% to 52%)
IFS
Day 5
2.40 (2.23 to 2.58)
1.81 (1.69 to 1.95)
-0.19 (-0.46 to 0.09)
51% (47% to 55%)
39% (32% to 45%)
GFS
Day 5
2.74 (2.59 to 2.90)
2.10 (1.98 to 2.23)
0.01 (-0.34 to 0.33)
44% (39% to 49%)
30% (22% to 37%)
Persistence
Day 5
4.92 (4.52 to 5.29)
3.90 (3.56 to 4.23)
0.13 (-0.56 to 0.80)
−25% (−39% to −12%)
Climatology
Day 5
3.94 (3.48 to 4.38)
3.10 (2.72 to 3.51)
0.32 (-0.59 to 1.24)
20% (11% to 28%)
AIFS
Day 7
3.00 (2.71 to 3.32)
2.26 (2.07 to 2.47)
0.13 (-0.19 to 0.46)
43% (36% to 49%)
24% (14% to 32%)
IFS
Day 7
3.42 (3.08 to 3.83)
2.54 (2.33 to 2.76)
-0.24 (-0.62 to 0.12)
35% (27% to 42%)
13% (2% to 23%)
GFS
Day 7
3.47 (3.15 to 3.81)
2.67 (2.44 to 2.93)
0.10 (-0.30 to 0.49)
34% (26% to 42%)
12% (−1% to 22%)
Persistence
Day 7
5.26 (4.76 to 5.74)
4.14 (3.72 to 4.55)
0.19 (-0.68 to 1.04)
−33% (−47% to −21%)
Climatology
Day 7
3.94 (3.48 to 4.37)
3.10 (2.72 to 3.50)
0.32 (-0.57 to 1.25)
25% (17% to 32%)
AIFS
Day 10
4.15 (3.79 to 4.51)
3.21 (2.94 to 3.50)
-0.11 (-0.65 to 0.38)
20% (11% to 28%)
−6% (−21% to 8%)
IFS
Day 10
4.57 (4.17 to 4.99)
3.53 (3.24 to 3.85)
-0.46 (-1.10 to 0.14)
12% (4% to 20%)
−16% (−30% to −4%)
GFS
Day 10
4.51 (4.17 to 4.83)
3.55 (3.30 to 3.80)
0.06 (-0.55 to 0.62)
14% (5% to 22%)
−15% (−29% to −2%)
Persistence
Day 10
5.22 (4.71 to 5.71)
4.11 (3.69 to 4.54)
0.21 (-0.88 to 1.28)
−33% (−47% to −20%)
Climatology
Day 10
3.94 (3.48 to 4.36)
3.10 (2.72 to 3.50)
0.32 (-0.58 to 1.24)
25% (17% to 32%)
Persistence's own row and the skill against it rest on 248 days, 7 September 2025 to 12 May 2026, on the 14,636 of 14,824 station-times where persistence can be scored as well. Every other score needs no earlier measurement.
All scores: 2 m temperature, bias-correctedRMSE, MAE, bias and skill for every model and lead
°C, with 95% intervals in brackets
Series
Lead
RMSE
MAE
Bias
Skill against persistence
Skill against climatology
AIFS
Day 1
1.17 (1.11 to 1.23)
0.87 (0.83 to 0.92)
0.02 (-0.04 to 0.08)
63% (60% to 65%)
72% (68% to 75%)
IFS
Day 1
1.42 (1.33 to 1.51)
1.06 (1.00 to 1.14)
0.00 (-0.12 to 0.12)
55% (51% to 58%)
66% (62% to 69%)
GFS
Day 1
1.76 (1.64 to 1.89)
1.33 (1.25 to 1.43)
-0.06 (-0.26 to 0.14)
44% (38% to 49%)
58% (52% to 62%)
Persistence
Day 1
3.13 (2.91 to 3.34)
2.40 (2.24 to 2.57)
0.01 (-0.18 to 0.19)
25% (15% to 33%)
Climatology
Day 1
4.17 (3.62 to 4.68)
3.26 (2.81 to 3.74)
-0.13 (-1.13 to 0.90)
−33% (−49% to −18%)
AIFS
Day 3
1.48 (1.40 to 1.55)
1.12 (1.06 to 1.18)
0.02 (-0.08 to 0.12)
68% (65% to 71%)
66% (60% to 69%)
IFS
Day 3
1.76 (1.65 to 1.88)
1.34 (1.25 to 1.43)
-0.01 (-0.16 to 0.14)
62% (59% to 65%)
59% (53% to 63%)
GFS
Day 3
2.12 (2.01 to 2.25)
1.63 (1.55 to 1.73)
-0.09 (-0.33 to 0.16)
55% (51% to 58%)
51% (43% to 56%)
Persistence
Day 3
4.69 (4.32 to 5.05)
3.67 (3.37 to 3.98)
-0.06 (-0.56 to 0.44)
−9% (−24% to 3%)
Climatology
Day 3
4.29 (3.71 to 4.85)
3.36 (2.86 to 3.89)
-0.15 (-1.20 to 0.94)
8% (−4% to 19%)
AIFS
Day 5
2.14 (2.02 to 2.26)
1.62 (1.53 to 1.71)
0.01 (-0.21 to 0.21)
59% (56% to 62%)
51% (44% to 56%)
IFS
Day 5
2.39 (2.20 to 2.60)
1.82 (1.68 to 1.98)
-0.04 (-0.32 to 0.21)
54% (50% to 58%)
45% (37% to 51%)
GFS
Day 5
2.77 (2.58 to 2.96)
2.13 (1.99 to 2.28)
-0.12 (-0.41 to 0.18)
47% (42% to 52%)
36% (26% to 44%)
Persistence
Day 5
5.20 (4.75 to 5.63)
4.14 (3.76 to 4.52)
-0.14 (-0.91 to 0.63)
−19% (−36% to −6%)
Climatology
Day 5
4.35 (3.74 to 4.94)
3.41 (2.89 to 3.97)
-0.16 (-1.28 to 0.96)
16% (5% to 26%)
AIFS
Day 7
3.11 (2.81 to 3.45)
2.35 (2.15 to 2.57)
-0.01 (-0.43 to 0.40)
44% (37% to 50%)
29% (17% to 38%)
IFS
Day 7
3.45 (3.07 to 3.90)
2.55 (2.34 to 2.80)
-0.10 (-0.51 to 0.30)
38% (30% to 45%)
21% (8% to 32%)
GFS
Day 7
3.53 (3.18 to 3.91)
2.72 (2.48 to 2.98)
-0.11 (-0.50 to 0.27)
36% (28% to 44%)
19% (5% to 31%)
Persistence
Day 7
5.55 (5.00 to 6.08)
4.37 (3.89 to 4.85)
-0.23 (-1.20 to 0.75)
−27% (−44% to −13%)
Climatology
Day 7
4.36 (3.74 to 4.97)
3.42 (2.90 to 3.98)
-0.18 (-1.35 to 1.01)
21% (12% to 31%)
AIFS
Day 10
4.29 (3.89 to 4.69)
3.33 (3.01 to 3.67)
-0.03 (-0.73 to 0.71)
21% (11% to 31%)
1% (−18% to 16%)
IFS
Day 10
4.65 (4.28 to 5.02)
3.61 (3.33 to 3.90)
-0.21 (-0.91 to 0.48)
15% (7% to 21%)
−7% (−25% to 7%)
GFS
Day 10
4.51 (4.16 to 4.84)
3.53 (3.27 to 3.78)
-0.30 (-0.87 to 0.27)
18% (8% to 26%)
−4% (−22% to 10%)
Persistence
Day 10
5.47 (4.92 to 5.98)
4.37 (3.89 to 4.85)
-0.43 (-1.63 to 0.81)
−26% (−44% to −11%)
Climatology
Day 10
4.34 (3.69 to 4.99)
3.39 (2.86 to 3.98)
-0.20 (-1.42 to 1.03)
21% (10% to 31%)
Persistence's own row and the skill against it rest on 218 days, 7 October 2025 to 12 May 2026, on the 12,838 of 13,028 station-times where persistence can be scored as well. Every other score needs no earlier measurement.
All scores: 10 m wind speed, rawRMSE, MAE, bias and skill for every model and lead
m/s, with 95% intervals in brackets
Series
Lead
RMSE
MAE
Bias
Skill against persistence
Skill against climatology
AIFS
Day 1
1.23 (1.18 to 1.29)
0.94 (0.90 to 0.98)
-0.30 (-0.35 to -0.25)
44% (41% to 47%)
38% (35% to 42%)
IFS
Day 1
1.23 (1.18 to 1.28)
0.93 (0.90 to 0.97)
-0.19 (-0.24 to -0.13)
44% (40% to 47%)
38% (34% to 42%)
GFS
Day 1
1.28 (1.22 to 1.34)
0.97 (0.93 to 1.01)
-0.00 (-0.07 to 0.07)
42% (38% to 45%)
36% (32% to 39%)
Persistence
Day 1
2.20 (2.04 to 2.35)
1.63 (1.51 to 1.75)
-0.01 (-0.09 to 0.06)
−10% (−17% to −4%)
Climatology
Day 1
1.99 (1.84 to 2.15)
1.58 (1.48 to 1.69)
0.35 (0.02 to 0.63)
9% (4% to 15%)
AIFS
Day 3
1.41 (1.34 to 1.47)
1.06 (1.01 to 1.11)
-0.38 (-0.45 to -0.30)
47% (43% to 50%)
29% (25% to 33%)
IFS
Day 3
1.43 (1.37 to 1.49)
1.07 (1.03 to 1.12)
-0.23 (-0.29 to -0.17)
46% (42% to 49%)
28% (24% to 32%)
GFS
Day 3
1.53 (1.45 to 1.61)
1.15 (1.09 to 1.21)
-0.07 (-0.14 to 0.01)
42% (38% to 46%)
23% (18% to 28%)
Persistence
Day 3
2.64 (2.44 to 2.85)
2.00 (1.84 to 2.18)
-0.01 (-0.18 to 0.16)
−33% (−41% to −24%)
Climatology
Day 3
1.99 (1.84 to 2.15)
1.58 (1.48 to 1.69)
0.35 (0.02 to 0.63)
25% (19% to 29%)
AIFS
Day 5
1.70 (1.58 to 1.81)
1.26 (1.18 to 1.35)
-0.48 (-0.60 to -0.36)
37% (32% to 41%)
15% (9% to 20%)
IFS
Day 5
1.77 (1.68 to 1.86)
1.32 (1.25 to 1.39)
-0.26 (-0.35 to -0.17)
34% (30% to 38%)
11% (6% to 16%)
GFS
Day 5
1.89 (1.77 to 2.00)
1.42 (1.33 to 1.51)
-0.07 (-0.18 to 0.04)
30% (25% to 34%)
5% (−1% to 12%)
Persistence
Day 5
2.68 (2.45 to 2.92)
2.01 (1.84 to 2.21)
-0.00 (-0.23 to 0.22)
−35% (−43% to −27%)
Climatology
Day 5
1.99 (1.84 to 2.15)
1.58 (1.48 to 1.69)
0.35 (0.03 to 0.63)
26% (21% to 30%)
AIFS
Day 7
2.10 (1.94 to 2.27)
1.57 (1.45 to 1.69)
-0.47 (-0.64 to -0.29)
24% (19% to 28%)
−6% (−13% to 1%)
IFS
Day 7
2.14 (2.02 to 2.27)
1.62 (1.52 to 1.72)
-0.23 (-0.38 to -0.07)
22% (17% to 27%)
−8% (−15% to −1%)
GFS
Day 7
2.31 (2.16 to 2.47)
1.75 (1.63 to 1.87)
0.01 (-0.17 to 0.20)
16% (10% to 22%)
−16% (−24% to −9%)
Persistence
Day 7
2.76 (2.54 to 2.98)
2.08 (1.91 to 2.26)
-0.01 (-0.28 to 0.26)
−38% (−46% to −31%)
Climatology
Day 7
1.99 (1.85 to 2.15)
1.58 (1.48 to 1.69)
0.35 (0.02 to 0.63)
28% (24% to 32%)
AIFS
Day 10
2.36 (2.19 to 2.54)
1.79 (1.66 to 1.93)
-0.42 (-0.63 to -0.22)
14% (6% to 22%)
−19% (−26% to −12%)
IFS
Day 10
2.56 (2.42 to 2.71)
1.94 (1.83 to 2.04)
-0.18 (-0.41 to 0.04)
7% (−2% to 14%)
−29% (−38% to −21%)
GFS
Day 10
2.82 (2.60 to 3.04)
2.12 (1.96 to 2.29)
0.28 (-0.01 to 0.56)
−3% (−15% to 7%)
−42% (−56% to −29%)
Persistence
Day 10
2.74 (2.51 to 3.00)
2.07 (1.90 to 2.28)
-0.01 (-0.33 to 0.33)
−38% (−48% to −29%)
Climatology
Day 10
1.99 (1.84 to 2.15)
1.58 (1.48 to 1.70)
0.35 (0.02 to 0.63)
27% (23% to 32%)
Persistence's own row and the skill against it rest on 248 days, 7 September 2025 to 12 May 2026, on the 14,623 of 14,824 station-times where persistence can be scored as well. Every other score needs no earlier measurement.
All scores: 10 m wind speed, bias-correctedRMSE, MAE, bias and skill for every model and lead
m/s, with 95% intervals in brackets
Series
Lead
RMSE
MAE
Bias
Skill against persistence
Skill against climatology
AIFS
Day 1
1.08 (1.01 to 1.15)
0.82 (0.78 to 0.86)
-0.02 (-0.06 to 0.02)
51% (48% to 55%)
46% (42% to 50%)
IFS
Day 1
1.08 (1.02 to 1.14)
0.82 (0.78 to 0.86)
-0.02 (-0.06 to 0.02)
51% (48% to 55%)
46% (42% to 50%)
GFS
Day 1
1.18 (1.11 to 1.26)
0.89 (0.84 to 0.94)
-0.02 (-0.07 to 0.04)
47% (43% to 50%)
41% (36% to 45%)
Persistence
Day 1
2.22 (2.05 to 2.39)
1.65 (1.51 to 1.79)
0.00 (-0.08 to 0.08)
−11% (−19% to −4%)
Climatology
Day 1
2.00 (1.81 to 2.20)
1.54 (1.42 to 1.69)
0.05 (-0.24 to 0.33)
10% (4% to 16%)
AIFS
Day 3
1.27 (1.19 to 1.34)
0.95 (0.90 to 1.00)
-0.03 (-0.10 to 0.03)
52% (49% to 55%)
37% (32% to 41%)
IFS
Day 3
1.32 (1.25 to 1.40)
0.99 (0.94 to 1.04)
-0.01 (-0.07 to 0.05)
50% (47% to 53%)
34% (29% to 39%)
GFS
Day 3
1.47 (1.37 to 1.56)
1.10 (1.03 to 1.17)
-0.01 (-0.09 to 0.08)
45% (40% to 49%)
27% (20% to 33%)
Persistence
Day 3
2.67 (2.44 to 2.90)
2.02 (1.83 to 2.21)
0.02 (-0.18 to 0.22)
−32% (−42% to −23%)
Climatology
Day 3
2.01 (1.82 to 2.22)
1.56 (1.43 to 1.70)
0.06 (-0.24 to 0.35)
24% (19% to 29%)
AIFS
Day 5
1.58 (1.47 to 1.71)
1.19 (1.11 to 1.28)
-0.04 (-0.19 to 0.10)
41% (37% to 46%)
21% (15% to 28%)
IFS
Day 5
1.69 (1.59 to 1.80)
1.27 (1.20 to 1.34)
0.01 (-0.12 to 0.13)
37% (33% to 41%)
16% (10% to 22%)
GFS
Day 5
1.86 (1.73 to 1.99)
1.39 (1.29 to 1.50)
-0.02 (-0.13 to 0.10)
31% (26% to 36%)
8% (0.3% to 15.2%)
Persistence
Day 5
2.71 (2.44 to 2.99)
2.03 (1.84 to 2.26)
0.02 (-0.26 to 0.30)
−34% (−43% to −26%)
Climatology
Day 5
2.02 (1.83 to 2.23)
1.56 (1.43 to 1.71)
0.07 (-0.24 to 0.36)
25% (21% to 30%)
AIFS
Day 7
2.04 (1.84 to 2.24)
1.53 (1.39 to 1.68)
-0.01 (-0.24 to 0.21)
28% (21% to 34%)
−1% (−11% to 8%)
IFS
Day 7
2.11 (1.96 to 2.26)
1.60 (1.48 to 1.72)
0.04 (-0.15 to 0.24)
25% (18% to 32%)
−4% (−14% to 5%)
GFS
Day 7
2.29 (2.12 to 2.47)
1.74 (1.60 to 1.87)
0.05 (-0.12 to 0.23)
19% (11% to 26%)
−14% (−23% to −4%)
Persistence
Day 7
2.82 (2.56 to 3.10)
2.14 (1.94 to 2.36)
0.01 (-0.32 to 0.37)
−40% (−47% to −33%)
Climatology
Day 7
2.02 (1.83 to 2.22)
1.56 (1.44 to 1.71)
0.08 (-0.23 to 0.38)
28% (25% to 32%)
AIFS
Day 10
2.35 (2.16 to 2.55)
1.79 (1.64 to 1.95)
0.01 (-0.27 to 0.27)
17% (8% to 26%)
−17% (−24% to −8%)
IFS
Day 10
2.59 (2.42 to 2.75)
1.96 (1.84 to 2.09)
0.01 (-0.30 to 0.33)
9% (−2% to 18%)
−29% (−38% to −20%)
GFS
Day 10
2.89 (2.64 to 3.13)
2.18 (2.00 to 2.37)
0.05 (-0.34 to 0.44)
−2% (−15% to 10%)
−43% (−56% to −31%)
Persistence
Day 10
2.84 (2.56 to 3.14)
2.15 (1.95 to 2.39)
0.02 (-0.38 to 0.43)
−41% (−52% to −31%)
Climatology
Day 10
2.02 (1.84 to 2.21)
1.56 (1.45 to 1.70)
0.09 (-0.23 to 0.41)
29% (23% to 34%)
Persistence's own row and the skill against it rest on 218 days, 7 October 2025 to 12 May 2026, on the 12,813 of 13,028 station-times where persistence can be scored as well. Every other score needs no earlier measurement.
All scores: solar energy, rawRMSE, MAE, bias and skill for every model and lead
kWh/m²/day, with 95% intervals in brackets
Series
Lead
RMSE
MAE
Bias
Skill against persistence
Skill against climatology
AIFS
Day 1
0.46 (0.37 to 0.55)
0.34 (0.27 to 0.41)
0.02 (-0.03 to 0.07)
66% (62% to 69%)
57% (50% to 62%)
IFS
Day 1
0.51 (0.42 to 0.60)
0.36 (0.30 to 0.43)
0.16 (0.12 to 0.20)
63% (56% to 66%)
53% (43% to 59%)
GFS
Day 1
0.63 (0.53 to 0.72)
0.44 (0.37 to 0.51)
0.08 (0.03 to 0.13)
55% (44% to 60%)
43% (27% to 52%)
Persistence
Day 1
1.41 (1.02 to 1.75)
0.98 (0.70 to 1.27)
0.00 (-0.10 to 0.11)
−25% (−39% to −12%)
Climatology
Day 1
1.09 (0.77 to 1.39)
0.81 (0.57 to 1.09)
-0.10 (-0.28 to 0.07)
20% (11% to 28%)
AIFS
Day 3
0.60 (0.46 to 0.73)
0.43 (0.33 to 0.53)
0.03 (-0.03 to 0.08)
59% (54% to 62%)
45% (38% to 50%)
IFS
Day 3
0.66 (0.53 to 0.80)
0.46 (0.37 to 0.56)
0.17 (0.12 to 0.21)
55% (48% to 59%)
39% (29% to 47%)
GFS
Day 3
0.80 (0.64 to 0.96)
0.55 (0.44 to 0.67)
0.10 (0.01 to 0.17)
45% (36% to 51%)
26% (12% to 35%)
Persistence
Day 3
1.50 (1.08 to 1.89)
1.06 (0.76 to 1.38)
-0.02 (-0.19 to 0.14)
−34% (−44% to −22%)
Climatology
Day 3
1.09 (0.77 to 1.39)
0.81 (0.57 to 1.09)
-0.10 (-0.29 to 0.07)
25% (18% to 31%)
AIFS
Day 5
0.77 (0.58 to 0.94)
0.55 (0.41 to 0.69)
0.01 (-0.07 to 0.08)
48% (44% to 52%)
30% (22% to 36%)
IFS
Day 5
0.86 (0.66 to 1.04)
0.60 (0.46 to 0.74)
0.15 (0.06 to 0.22)
42% (35% to 47%)
21% (10% to 30%)
GFS
Day 5
1.02 (0.78 to 1.24)
0.71 (0.54 to 0.88)
0.12 (0.01 to 0.24)
31% (22% to 37%)
6% (−9% to 16%)
Persistence
Day 5
1.52 (1.10 to 1.89)
1.08 (0.77 to 1.40)
-0.03 (-0.25 to 0.19)
−35% (−47% to −25%)
Climatology
Day 5
1.09 (0.77 to 1.39)
0.81 (0.57 to 1.09)
-0.10 (-0.29 to 0.07)
26% (20% to 32%)
AIFS
Day 7
0.98 (0.72 to 1.22)
0.69 (0.51 to 0.89)
0.01 (-0.11 to 0.13)
35% (30% to 40%)
10% (−1% to 18%)
GFS
Day 7
1.22 (0.87 to 1.54)
0.83 (0.61 to 1.09)
0.05 (-0.09 to 0.20)
20% (10% to 27%)
−11% (−24.5% to −0.4%)
Persistence
Day 7
1.56 (1.13 to 1.94)
1.12 (0.80 to 1.45)
-0.02 (-0.30 to 0.26)
−39% (−50% to −30%)
Climatology
Day 7
1.09 (0.77 to 1.39)
0.81 (0.57 to 1.09)
-0.10 (-0.29 to 0.07)
28% (23% to 33%)
AIFS
Day 10
1.13 (0.81 to 1.42)
0.81 (0.58 to 1.06)
-0.01 (-0.17 to 0.14)
29% (21% to 36%)
−4% (−12% to 3%)
GFS
Day 10
1.36 (0.97 to 1.72)
0.96 (0.69 to 1.26)
-0.05 (-0.25 to 0.12)
15% (6% to 21%)
−25% (−32% to −18%)
Persistence
Day 10
1.65 (1.16 to 2.08)
1.18 (0.82 to 1.56)
-0.05 (-0.39 to 0.28)
−46% (−55% to −37%)
Climatology
Day 10
1.09 (0.77 to 1.39)
0.81 (0.57 to 1.09)
-0.10 (-0.28 to 0.07)
32% (27% to 36%)
Persistence's own row and the skill against it rest on 230 days, 7 September 2025 to 12 May 2026: those of the 247 days scored on which persistence can be scored as well. Every other score needs no earlier measurement.
All scores: solar energy, bias-correctedRMSE, MAE, bias and skill for every model and lead
kWh/m²/day, with 95% intervals in brackets
Series
Lead
RMSE
MAE
Bias
Skill against persistence
Skill against climatology
AIFS
Day 1
0.46 (0.35 to 0.56)
0.33 (0.26 to 0.41)
-0.01 (-0.05 to 0.04)
66% (61% to 69%)
58% (50% to 62%)
IFS
Day 1
0.48 (0.37 to 0.57)
0.34 (0.27 to 0.42)
-0.01 (-0.05 to 0.04)
65% (59% to 68%)
56% (47% to 61%)
GFS
Day 1
0.61 (0.49 to 0.72)
0.43 (0.36 to 0.51)
0.01 (-0.06 to 0.07)
55% (45% to 61%)
44% (27% to 52%)
Persistence
Day 1
1.40 (0.95 to 1.78)
0.95 (0.65 to 1.28)
0.01 (-0.09 to 0.11)
−24% (−38% to −11%)
Climatology
Day 1
1.08 (0.72 to 1.41)
0.79 (0.53 to 1.09)
-0.05 (-0.23 to 0.12)
19% (10% to 27%)
AIFS
Day 3
0.59 (0.43 to 0.73)
0.42 (0.31 to 0.53)
-0.00 (-0.07 to 0.05)
59% (53% to 63%)
46% (39% to 51%)
IFS
Day 3
0.63 (0.47 to 0.79)
0.44 (0.34 to 0.55)
-0.01 (-0.06 to 0.05)
56% (49% to 61%)
42% (32% to 49%)
GFS
Day 3
0.78 (0.58 to 0.96)
0.54 (0.42 to 0.66)
-0.02 (-0.11 to 0.07)
46% (36% to 52%)
29% (15% to 37%)
Persistence
Day 3
1.50 (0.99 to 1.95)
1.03 (0.71 to 1.39)
-0.03 (-0.18 to 0.15)
−32% (−42% to −20%)
Climatology
Day 3
1.10 (0.73 to 1.43)
0.80 (0.54 to 1.10)
-0.06 (-0.25 to 0.12)
24% (17% to 30%)
AIFS
Day 5
0.77 (0.55 to 0.96)
0.54 (0.39 to 0.70)
-0.01 (-0.10 to 0.07)
47% (40% to 51%)
30% (22% to 36%)
IFS
Day 5
0.83 (0.60 to 1.04)
0.57 (0.42 to 0.74)
-0.02 (-0.10 to 0.06)
42% (35% to 48%)
25% (14% to 32%)
GFS
Day 5
1.01 (0.73 to 1.27)
0.69 (0.51 to 0.90)
-0.04 (-0.16 to 0.07)
30% (19% to 37%)
8% (−5% to 17%)
Persistence
Day 5
1.50 (0.98 to 1.95)
1.02 (0.71 to 1.37)
-0.03 (-0.26 to 0.22)
−31% (−42% to −19%)
Climatology
Day 5
1.10 (0.73 to 1.44)
0.80 (0.54 to 1.11)
-0.06 (-0.26 to 0.12)
23% (16% to 29%)
AIFS
Day 7
0.98 (0.68 to 1.25)
0.68 (0.48 to 0.90)
-0.05 (-0.17 to 0.08)
36% (29% to 42%)
11% (1% to 20%)
GFS
Day 7
1.22 (0.82 to 1.60)
0.83 (0.57 to 1.12)
-0.04 (-0.18 to 0.12)
20% (7% to 30%)
−11% (−27% to 3%)
Persistence
Day 7
1.59 (1.05 to 2.06)
1.11 (0.75 to 1.50)
-0.02 (-0.30 to 0.30)
−38% (−48% to −28%)
Climatology
Day 7
1.10 (0.74 to 1.44)
0.81 (0.55 to 1.11)
-0.07 (-0.27 to 0.12)
28% (22% to 33%)
AIFS
Day 10
1.12 (0.76 to 1.45)
0.78 (0.54 to 1.06)
-0.09 (-0.23 to 0.05)
29% (17% to 38%)
−2% (−11% to 5%)
GFS
Day 10
1.34 (0.91 to 1.73)
0.93 (0.64 to 1.26)
-0.10 (-0.26 to 0.05)
15% (4% to 23%)
−22% (−29% to −15%)
Persistence
Day 10
1.64 (1.06 to 2.16)
1.14 (0.76 to 1.56)
-0.09 (-0.40 to 0.27)
−44% (−57% to −30%)
Climatology
Day 10
1.10 (0.74 to 1.42)
0.80 (0.54 to 1.10)
-0.08 (-0.28 to 0.10)
31% (23% to 36%)
Persistence's own row and the skill against it rest on 197 days, 9 October 2025 to 12 May 2026: those of the 216 days scored on which persistence can be scored as well. Every other score needs no earlier measurement.
Summer and autumn 2026: 174 days from 23 May to 12 November 2026, 9 of them in spring, at 30 German stations, raw forecasts. AIFS 2 had the lower RMSE against IFS at 3 of 5 leads for temperature, 1 of 5 leads for wind and 3 of 3 leads for solar; against GFS at 3 of 5 leads for temperature, 3 of 5 leads for wind and 4 of 5 leads for solar. Of the rest, 11 could not be separated. The 28 comparisons share days and stations, so they are not 28 independent results. IFS and GFS are scored at 0.25°, as openly distributed.
Skill compares each model with a simple reference: climatology (the average for that date and hour) or persistence (the latest measurement carried forward). In the violet area a model beats that reference; skill is not tested, so the area is not the green of a verdict. IFS and GFS are scored at 0.25° and ICON at 0.125°, as openly distributed, not at their native resolution. Each AIFS version has its own days and seasons, so do not compare the two.
How much data this rests onDays scored, seasons still missing, and the smallest gap the sample can detect
No AIFS version has been running for a full year yet, so every result on this page is provisional. A result would stop being provisional once a version covers at least 365 days and every season.
Days scored, period, seasons missing and effective days, by AIFS version and variable
Version
Variable
Days scored
Period
Seasons still missing
Effective days
AIFS 2
2 m temperature
174
23 May 2026 to 12 November 2026
winter, spring
about 37
AIFS 2
10 m wind speed
174
23 May 2026 to 12 November 2026
winter, spring
about 64
AIFS 2
Solar energy
174
23 May 2026 to 12 November 2026
winter, spring
about 51
AIFS 1.1
2 m temperature
248
7 September 2025 to 12 May 2026
summer
about 62
AIFS 1.1
10 m wind speed
248
7 September 2025 to 12 May 2026
summer
about 93
AIFS 1.1
Solar energy
247
7 September 2025 to 12 May 2026
summer
about 87
The direction of the differences was steadier than their size. These counts cover raw forecasts at the 30 German stations, both AIFS versions, every lead and model compared. The sign was the same in both halves of the period in 52 of 56 comparisons, and AIFS had the lower error in 378 of 420 monthly comparisons. The smallest RMSE difference a sample would detect 80% of the time, after the Holm correction, depends on the version and the lead. For AIFS 2, over 174 days and leads of 1 to 10 days, it is 0.10 to 1.04 °C for 2 m temperature, 0.06 to 0.30 m/s for 10 m wind and 0.10 to 0.30 kWh/m²/day for daily solar energy. For AIFS 1.1, over 248 days and leads of 1 to 10 days, it is 0.19 to 0.93 °C for 2 m temperature, 0.06 to 0.49 m/s for 10 m wind and 0.05 to 0.28 kWh/m²/day for daily solar energy. A smaller real gap may not show as a verdict.
AIGFS at the same stationsNOAA's operational AI model, in a set of its own
AIGFS is NOAA's operational AI model. It starts from the GFS analysis, not from ECMWF's, and it publishes no surface height, so its temperature is moved to station height from the GFS surface height. It is scored here in a set of its own with AIFS, IFS and GFS, at the 30 German stations in the AIFS 2 period: 156 days, 23 May 2026 to 4 November 2026, with no winter or spring. Its archive has no solar, so this block shows 2 m temperature and 10 m wind at 00:00 and 12:00 UTC. The Reference and Forecasts switches apply; the Version switch does not.
2 m temperature: skill against climatology
Raw forecasts, AIFS 2, 156 days, 23 May 2026 to 4 November 2026. Hover for values; AIFS shows its 95% range, others on hover.
2 m temperature: skill against persistence
Raw forecasts, AIFS 2, 156 days, 23 May 2026 to 4 November 2026, on the 9,181 of 9,325 station-times where persistence can be scored as well. Hover for values; AIFS shows its 95% range, others on hover.
2 m temperature: skill against climatology
Bias-corrected forecasts, AIFS 2, 126 days, 22 June 2026 to 4 November 2026. Hover for values; AIFS shows its 95% range, others on hover.
2 m temperature: skill against persistence
Bias-corrected forecasts, AIFS 2, 126 days, 22 June 2026 to 4 November 2026, on the 7,396 of 7,526 station-times where persistence can be scored as well. Hover for values; AIFS shows its 95% range, others on hover.
2 m temperature, AIFS 2: how AIFS's error compared, over 156 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
▼−10%, AIFS lower error, tested
▼−12%, AIFS lower error, tested
▼−12%, AIFS lower error, tested
−9%, not separated
−11%, not separated
GFS
▼−28%, AIFS lower error, tested
▼−28%, AIFS lower error, tested
▼−24%, AIFS lower error, tested
−20%, not separated
−16%, not separated
AIGFS
▼−16%, AIFS lower error, tested
▼−16%, AIFS lower error, tested
▼−13%, AIFS lower error, tested
−11%, not separated
−14%, not separated
Each number is AIFS's RMSE minus the other model's, as a percentage of the other model's RMSE: negative means AIFS's was lower. Only comparisons with AIFS are tested, so the card does not say how AIGFS compares with GFS.
2 m temperature, AIFS 2: how AIFS's error compared, over 126 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
−13%, not separated
−14%, not separated
−13%, not separated
−5%, not separated
−14%, not separated
GFS
−29%, not separated
−27%, not separated
−23%, not separated
−20%, not separated
−18%, not separated
AIGFS
−17%, not separated
−16%, not separated
−12%, not separated
−10%, not separated
−13%, not separated
Each number is AIFS's RMSE minus the other model's, as a percentage of the other model's RMSE: negative means AIFS's was lower. Only comparisons with AIFS are tested, so the card does not say how AIGFS compares with GFS. This sample holds 12.6 blocks of 10 days, fewer than the 15 a verdict needs, so every cell is grey.
10 m wind speed: skill against climatology
Raw forecasts, AIFS 2, 156 days, 23 May 2026 to 4 November 2026. Hover for values; AIFS shows its 95% range, others on hover.
10 m wind speed: skill against persistence
Raw forecasts, AIFS 2, 156 days, 23 May 2026 to 4 November 2026, on the 9,031 of 9,283 station-times where persistence can be scored as well. Hover for values; AIFS shows its 95% range, others on hover.
10 m wind speed: skill against climatology
Bias-corrected forecasts, AIFS 2, 126 days, 22 June 2026 to 4 November 2026. Hover for values; AIFS shows its 95% range, others on hover.
10 m wind speed: skill against persistence
Bias-corrected forecasts, AIFS 2, 126 days, 22 June 2026 to 4 November 2026, on the 7,209 of 7,435 station-times where persistence can be scored as well. Hover for values; AIFS shows its 95% range, others on hover.
10 m wind speed, AIFS 2: how AIFS's error compared, over 156 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
−1%, not separated
−3%, not separated
▼−7%, AIFS lower error, tested
−5%, not separated
0%, not separated
GFS
−4%, not separated
▼−7%, AIFS lower error, tested
▼−11%, AIFS lower error, tested
▼−9%, AIFS lower error, tested
−3%, not separated
AIGFS
−2%, not separated
−5%, not separated
−8%, not separated
−6%, not separated
0%, not separated
Each number is AIFS's RMSE minus the other model's, as a percentage of the other model's RMSE: negative means AIFS's was lower. Only comparisons with AIFS are tested, so the card does not say how AIGFS compares with GFS.
10 m wind speed, AIFS 2: how AIFS's error compared, over 126 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
−2%, not separated
−6%, not separated
−9%, not separated
−7%, not separated
0%, not separated
GFS
−10%, not separated
−13%, not separated
−15%, not separated
−14%, not separated
−4%, not separated
AIGFS
−6%, not separated
−10%, not separated
−11%, not separated
−7%, not separated
+1%, not separated
Each number is AIFS's RMSE minus the other model's, as a percentage of the other model's RMSE: negative means AIFS's was lower. Only comparisons with AIFS are tested, so the card does not say how AIGFS compares with GFS. This sample holds 12.6 blocks of 10 days, fewer than the 15 a verdict needs, so every cell is grey.
▼AIFS lower error, tested (better)
▲AIFS higher error, tested (worse)
Not separated: no clear difference, too few days, or the verdict depends on how the station is read or on the height correction (hover says which)
Bias, activity and events, with AIGFS
AIFS 2, the same 156 days. Bold and underlined marks the value nearest the ideal at each lead, as printed. It is not a verdict: these values are not tested against each other.
2 m temperature: bias, in °CForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
+0.10
+0.12
+0.27
+0.50
−0.10
IFS
+0.11
+0.14
+0.39
+0.58
+0.49
GFS
−0.21
−0.16
−0.07
+0.00
−0.20
AIGFS
−0.19
−0.30
−0.32
−0.23
−0.58
2 m temperature: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.98
0.98
1.00
0.99
0.94
IFS
0.95
0.95
0.97
0.96
0.99
GFS
0.95
0.96
0.96
0.97
0.98
AIGFS
0.94
0.93
0.95
0.95
0.95
2 m temperature of 0 °C or belowObserved 1 times on 1 days in 9325 station-times (0.0%). Too few events to score: at least 30, on at least 10 days, are needed.
2 m temperature: bias, in °CForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
−0.07
−0.06
−0.08
−0.23
−0.08
IFS
−0.01
+0.02
+0.03
−0.00
−0.17
GFS
+0.13
+0.17
+0.24
+0.21
+0.09
AIGFS
+0.10
+0.16
+0.13
+0.08
+0.15
2 m temperature: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.95
0.95
0.97
0.96
0.94
IFS
0.94
0.95
0.99
1.00
0.99
GFS
0.96
0.98
0.98
1.00
0.99
AIGFS
0.94
0.93
0.93
0.96
0.95
2 m temperature of 0 °C or belowObserved 1 times on 1 days in 7526 station-times (0.0%). Too few events to score: at least 30, on at least 10 days, are needed.
10 m wind: bias, in m/sForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
−0.26
−0.35
−0.38
−0.47
−0.41
IFS
−0.21
−0.27
−0.27
−0.35
−0.42
GFS
+0.04
−0.04
−0.05
−0.09
−0.11
AIGFS
−0.01
−0.12
−0.23
−0.34
−0.24
10 m wind: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.92
0.89
0.85
0.82
0.86
IFS
0.95
0.94
0.92
0.87
0.85
GFS
1.01
0.97
0.92
0.94
0.94
AIGFS
1.00
0.95
0.88
0.87
0.87
10 m wind of 8 m/s or more: frequency biasObserved 274 times on 71 days in 9283 station-times (3%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.47
0.38
0.34
0.26
0.37
IFS
0.62
0.64
0.57
0.48
0.43
GFS
1.11
1.02
0.88
0.91
0.80
AIGFS
0.85
0.65
0.46
0.34
0.50
10 m wind of 8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.30
0.23
0.18
0.12
0.05
IFS
0.34
0.30
0.20
0.13
0.07
GFS
0.36
0.27
0.21
0.14
0.10
AIGFS
0.33
0.23
0.13
0.09
0.04
10 m wind of 10.8 m/s or more: frequency biasObserved 64 times on 31 days in 9283 station-times (0.7%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.16
0.17
0.08
0.11
0.19
IFS
0.67
0.61
0.72
0.50
0.64
GFS
0.70
0.63
0.55
0.56
0.52
AIGFS
0.44
0.34
0.11
0.13
0.08
10 m wind of 10.8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.15
0.15
0.08
0.06
-0.00
IFS
0.30
0.24
0.17
0.06
0.08
GFS
0.34
0.27
0.12
0.06
0.10
AIGFS
0.31
0.19
0.06
0.01
0.01
10 m wind: bias, in m/sForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
+0.02
+0.02
+0.04
+0.04
−0.02
IFS
+0.01
+0.01
−0.03
+0.00
−0.06
GFS
+0.02
+0.01
+0.04
+0.01
−0.04
AIGFS
+0.02
+0.03
+0.06
−0.01
−0.08
10 m wind: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.92
0.90
0.86
0.85
0.86
IFS
0.95
0.94
0.93
0.89
0.82
GFS
1.01
0.97
0.94
0.98
0.92
AIGFS
1.00
0.96
0.90
0.89
0.86
10 m wind of 8 m/s or more: frequency biasObserved 226 times on 59 days in 7435 station-times (3%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.85
0.81
0.79
0.77
0.84
IFS
0.85
0.93
0.85
0.79
0.66
GFS
1.11
1.07
0.94
1.01
0.99
AIGFS
1.03
0.92
0.87
0.67
0.75
10 m wind of 8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.45
0.36
0.29
0.21
0.09
IFS
0.46
0.39
0.28
0.18
0.09
GFS
0.42
0.36
0.24
0.17
0.12
AIGFS
0.44
0.33
0.23
0.15
0.09
10 m wind of 10.8 m/s or more: frequency biasObserved 58 times on 28 days in 7435 station-times (0.8%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.41
0.33
0.22
0.26
0.34
IFS
0.60
0.57
0.69
0.47
0.57
GFS
0.62
0.60
0.64
0.66
0.60
AIGFS
0.48
0.47
0.33
0.26
0.24
10 m wind of 10.8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.34
0.28
0.14
0.07
0.02
IFS
0.31
0.30
0.19
0.07
0.08
GFS
0.32
0.25
0.14
0.08
0.12
AIGFS
0.36
0.29
0.13
0.04
0.04
Other AI models at the same stationsGraphCast, FourCastNet, Pangu-Weather and Aurora, as re-run by NOAA
GraphCast, FourCastNet v2-small (not the newer FourCastNet 3), Pangu-Weather and Aurora are NOAA's re-runs from its open AIWP archive, started from IFS analyses. They are not the developers' own forecasts. AIGFS, AIFS and IFS are shown on the same days; AIGFS has more days in its own block above. The archive has no solar and covers only part of the AIFS 2 period, so this panel shows 2 m temperature and 10 m wind at 00:00 and 12:00 UTC. It rests on 68 days, with no winter or spring, fewer than the 15 blocks a verdict needs, so it shows differences without verdicts. The Reference and Forecasts switches apply; the Version switch does not.
2 m temperature
Raw forecasts against climatology, AIFS 2. 68 days, 21 July 2026 to 4 November 2026. Bars are 95% intervals.
2 m temperature
Raw forecasts against persistence, AIFS 2. 68 days, 21 July 2026 to 4 November 2026, on the 3,988 of 4,065 station-times where persistence can be scored as well. Bars are 95% intervals.
2 m temperature
Bias-corrected forecasts against climatology, AIFS 2. 43 days, 20 August 2026 to 4 November 2026. Bars are 95% intervals.
2 m temperature
Bias-corrected forecasts against persistence, AIFS 2. 43 days, 20 August 2026 to 4 November 2026, on the 2,517 of 2,566 station-times where persistence can be scored as well. Bars are 95% intervals.
2 m temperature, AIFS 2: how AIFS's error compared, over 68 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
−11%, not separated
−14%, not separated
−14%, not separated
−2%, not separated
−13%, not separated
AIGFS
−16%, not separated
−15%, not separated
−10%, not separated
−10%, not separated
−17%, not separated
GraphCast
−4%, not separated
−5%, not separated
−1%, not separated
−2%, not separated
−5%, not separated
FourCastNet v2
−22%, not separated
−21%, not separated
−14%, not separated
−2%, not separated
−7%, not separated
Pangu
−9%, not separated
−11%, not separated
−10%, not separated
−11%, not separated
−9%, not separated
Aurora
−4%, not separated
−4%, not separated
0%, not separated
+9%, not separated
+11%, not separated
▼AIFS lower error, tested (better)
▲AIFS higher error, tested (worse)
Not separated: no clear difference, too few days, or the verdict depends on how the station is read or on the height correction (hover says which)
2 m temperature, AIFS 2: how AIFS's error compared, over 43 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
−14%, not separated
−16%, not separated
−11%, not separated
+8%, not separated
−12%, not separated
AIGFS
−19%, not separated
−14%, not separated
−8%, not separated
−1%, not separated
−9%, not separated
GraphCast
−7%, not separated
−7%, not separated
−1%, not separated
+6%, not separated
−1%, not separated
FourCastNet v2
−26%, not separated
−22%, not separated
−14%, not separated
+4%, not separated
−12%, not separated
Pangu
−11%, not separated
−14%, not separated
−11%, not separated
−5%, not separated
−11%, not separated
Aurora
−6%, not separated
−9%, not separated
−2%, not separated
+17%, not separated
+18%, not separated
▼AIFS lower error, tested (better)
▲AIFS higher error, tested (worse)
Not separated: no clear difference, too few days, or the verdict depends on how the station is read or on the height correction (hover says which)
10 m wind speed
Raw forecasts against climatology, AIFS 2. 68 days, 21 July 2026 to 4 November 2026. Bars are 95% intervals.
10 m wind speed
Raw forecasts against persistence, AIFS 2. 68 days, 21 July 2026 to 4 November 2026, on the 3,961 of 4,059 station-times where persistence can be scored as well. Bars are 95% intervals.
10 m wind speed
Bias-corrected forecasts against climatology, AIFS 2. 43 days, 20 August 2026 to 4 November 2026. Bars are 95% intervals.
10 m wind speed
Bias-corrected forecasts against persistence, AIFS 2. 43 days, 20 August 2026 to 4 November 2026, on the 2,486 of 2,557 station-times where persistence can be scored as well. Bars are 95% intervals.
10 m wind speed, AIFS 2: how AIFS's error compared, over 68 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
−3%, not separated
−4%, not separated
−6%, not separated
−5%, not separated
+5%, not separated
AIGFS
−2%, not separated
−2%, not separated
−6%, not separated
−7%, not separated
+4%, not separated
GraphCast
−2%, not separated
0%, not separated
−4%, not separated
0%, not separated
+13%, not separated
FourCastNet v2
−12%, not separated
−11%, not separated
−9%, not separated
−7%, not separated
−4%, not separated
Pangu
−4%, not separated
−4%, not separated
−6%, not separated
−7%, not separated
+6%, not separated
Aurora
+2%, not separated
+2%, not separated
−4%, not separated
−11%, not separated
−5%, not separated
▼AIFS lower error, tested (better)
▲AIFS higher error, tested (worse)
Not separated: no clear difference, too few days, or the verdict depends on how the station is read or on the height correction (hover says which)
10 m wind speed, AIFS 2: how AIFS's error compared, over 43 days
Compared with
Day 1
Day 3
Day 5
Day 7
Day 10
IFS
−4%, not separated
−7%, not separated
−8%, not separated
−11%, not separated
+9%, not separated
AIGFS
−6%, not separated
−8%, not separated
−10%, not separated
−11%, not separated
+6%, not separated
GraphCast
0%, not separated
−3%, not separated
−6%, not separated
−2%, not separated
+16%, not separated
FourCastNet v2
−14%, not separated
−14%, not separated
−8%, not separated
−3%, not separated
−1%, not separated
Pangu
−5%, not separated
−6%, not separated
−8%, not separated
−13%, not separated
+8%, not separated
Aurora
+1%, not separated
−1%, not separated
+4%, not separated
+6%, not separated
+33%, not separated
▼AIFS lower error, tested (better)
▲AIFS higher error, tested (worse)
Not separated: no clear difference, too few days, or the verdict depends on how the station is read or on the height correction (hover says which)
Bias, activity and events, other AI models
AIFS 2, the same 68 days. Bold and underlined marks the value nearest the ideal at each lead, as printed. It is not a verdict: these values are not tested against each other.
2 m temperature: bias, in °CForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
−0.01
−0.02
+0.01
+0.24
−0.02
IFS
+0.10
+0.17
+0.46
+0.76
+0.20
AIGFS
−0.16
−0.32
−0.38
−0.23
−0.69
GraphCast
+0.20
+0.04
0.00
+0.45
−0.35
FourCastNet v2
+0.24
−0.17
−0.44
−0.37
−0.24
Pangu
+0.24
+0.16
+0.17
+0.51
+0.07
Aurora
+0.24
+0.15
+0.18
+0.24
+0.43
2 m temperature: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.95
0.91
0.92
0.97
0.97
IFS
0.93
0.91
0.92
0.99
1.03
AIGFS
0.91
0.87
0.87
1.01
0.99
GraphCast
0.92
0.89
0.88
0.99
0.95
FourCastNet v2
0.85
0.76
0.75
0.78
0.81
Pangu
0.93
0.92
0.93
1.00
0.97
Aurora
0.93
0.87
0.81
0.79
0.71
2 m temperature of 0 °C or belowObserved 1 times on 1 days in 4065 station-times (0.0%). Too few events to score: at least 30, on at least 10 days, are needed.
2 m temperature: bias, in °CForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
−0.06
−0.08
−0.03
−0.22
−0.11
IFS
−0.09
−0.22
−0.17
−0.33
−0.24
AIGFS
+0.15
+0.13
+0.21
+0.06
+0.65
GraphCast
−0.08
−0.11
+0.04
−0.19
−0.24
FourCastNet v2
+0.10
+0.30
+0.58
+0.65
+0.70
Pangu
−0.04
+0.04
+0.38
+0.29
+0.17
Aurora
−0.02
+0.02
+0.14
+0.20
+0.03
2 m temperature: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.92
0.88
0.86
0.94
1.06
IFS
0.90
0.87
0.78
0.90
0.96
AIGFS
0.89
0.83
0.83
1.02
1.16
GraphCast
0.91
0.86
0.80
0.93
0.94
FourCastNet v2
0.84
0.76
0.78
0.79
1.06
Pangu
0.92
0.90
0.89
0.99
1.08
Aurora
0.89
0.81
0.71
0.73
0.73
2 m temperature of 0 °C or belowObserved 1 times on 1 days in 2566 station-times (0.0%). Too few events to score: at least 30, on at least 10 days, are needed.
10 m wind: bias, in m/sForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
−0.24
−0.36
−0.41
−0.50
−0.40
IFS
−0.21
−0.28
−0.33
−0.35
−0.50
AIGFS
−0.03
−0.20
−0.27
−0.31
−0.43
GraphCast
−0.21
−0.27
−0.28
−0.36
−0.31
FourCastNet v2
−0.13
−0.30
−0.38
−0.55
−0.61
Pangu
−0.20
−0.26
−0.22
−0.28
−0.46
Aurora
−0.19
−0.34
−0.68
−1.14
−1.56
10 m wind: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.92
0.89
0.85
0.86
0.93
IFS
0.95
0.94
0.94
0.91
0.82
AIGFS
0.99
0.94
0.91
0.94
0.84
GraphCast
0.91
0.90
0.89
0.87
0.78
FourCastNet v2
0.96
0.91
0.85
0.86
0.87
Pangu
0.90
0.87
0.89
0.88
0.91
Aurora
0.89
0.85
0.76
0.65
0.51
10 m wind of 8 m/s or more: frequency biasObserved 103 times on 34 days in 4059 station-times (3%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.41
0.31
0.27
0.23
0.52
IFS
0.59
0.61
0.64
0.59
0.38
AIGFS
0.87
0.71
0.56
0.39
0.32
GraphCast
0.44
0.35
0.38
0.32
0.28
FourCastNet v2
0.67
0.45
0.37
0.30
0.38
Pangu
0.45
0.40
0.48
0.52
0.59
Aurora
0.42
0.31
0.15
0.02
0.00
10 m wind of 8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.25
0.20
0.11
0.10
0.04
IFS
0.27
0.23
0.21
0.14
0.05
AIGFS
0.30
0.24
0.16
0.12
0.06
GraphCast
0.26
0.23
0.14
0.12
0.10
FourCastNet v2
0.27
0.18
0.13
0.07
0.01
Pangu
0.24
0.19
0.13
0.13
0.08
Aurora
0.30
0.23
0.10
0.00
0.00
10 m wind of 10.8 m/s or moreObserved 17 times on 11 days in 4059 station-times (0.4%). Too few events to score: at least 30, on at least 10 days, are needed.
10 m wind: bias, in m/sForecast minus measurement, on average. Plus: too high. Minus: too low. Zero is best.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
+0.02
+0.03
+0.06
+0.17
+0.24
IFS
−0.01
−0.02
−0.02
+0.11
−0.25
AIGFS
+0.09
+0.07
+0.14
+0.12
−0.07
GraphCast
+0.02
+0.03
+0.07
+0.17
+0.09
FourCastNet v2
+0.09
+0.08
+0.13
+0.08
+0.28
Pangu
+0.05
+0.01
+0.05
+0.23
+0.16
Aurora
+0.02
0.00
−0.02
+0.00
+0.04
10 m wind: activityVariability of the forecast divided by that of the measurements, both as departures from climatology, with each station centred on its own mean at each valid hour first. 1.00 is best. Below 1 the forecast is too smooth, which flatters RMSE.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.92
0.91
0.88
0.89
0.96
IFS
0.95
0.95
0.91
0.91
0.82
AIGFS
1.01
0.98
0.94
1.00
0.88
GraphCast
0.91
0.93
0.92
0.90
0.78
FourCastNet v2
0.97
0.94
0.89
0.85
0.93
Pangu
0.91
0.90
0.92
0.89
0.95
Aurora
0.90
0.87
0.76
0.65
0.54
10 m wind of 8 m/s or more: frequency biasObserved 78 times on 21 days in 2557 station-times (3%). Forecast events per observed event. 1.00 is best. Above 1 the model calls the event too often, below 1 too rarely.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.90
0.92
0.94
0.91
1.36
IFS
0.87
0.99
1.01
1.19
0.45
AIGFS
1.15
1.09
1.15
1.13
0.90
GraphCast
0.92
0.95
0.96
1.04
0.78
FourCastNet v2
0.99
0.95
0.86
0.81
1.22
Pangu
0.82
0.78
1.00
1.19
1.50
Aurora
0.87
0.80
0.54
0.37
0.40
10 m wind of 8 m/s or more: equitable threat scoreHow well forecast and observed events coincide, beyond chance. 1 is perfect, 0 is chance, higher is better.
Model
Day 1
Day 3
Day 5
Day 7
Day 10
AIFS
0.41
0.33
0.18
0.22
0.08
IFS
0.42
0.35
0.27
0.17
0.11
AIGFS
0.42
0.36
0.25
0.20
0.11
GraphCast
0.34
0.34
0.22
0.27
0.17
FourCastNet v2
0.37
0.32
0.30
0.18
0.05
Pangu
0.41
0.31
0.22
0.18
0.13
Aurora
0.39
0.32
0.23
0.18
0.14
10 m wind of 10.8 m/s or moreObserved 13 times on 7 days in 2557 station-times (0.5%). Too few events to score: at least 30, on at least 10 days, are needed.
The figures, written outThe headline numbers for AIFS 2, raw forecasts, as sentences
After bias correction, none of the 28 comparisons can be judged yet. That sample holds 14.4 blocks of 10 days; a verdict needs 15.
Three days ahead, AIFS 2's RMSE was 66% below persistence for temperature, 42% for wind and 60% for solar; against climatology, 61%, 24% and 47%. The 2015–2024 station climatology is not adjusted for warming and ran 0.9 K too cold over these days, which raises every model's temperature skill against it. By day 7, its temperature RMSE was still 29% below climatology's; its wind RMSE could not be separated from climatology's, with IFS and GFS scoring lower against climatology.
RMSE favours smooth forecasts. At day 10, AIFS 2 carried 65% of the measured variability in daily solar energy, against 86% for GFS. Strong wind and frost are scored as events in the fold "Bias, activity and events".
AIFS 2, 2 m temperature, day 3, over 174 days: skill against climatology 61% (55% to 65%) and against persistence 66% (62% to 68%). Month by month, skill against climatology ranged from 52% to 70%. On the 144 days that can be bias-corrected, the correction moves RMSE from 1.60 to 1.49 °C.
AIFS 2, 10 m wind speed, day 3, over 174 days: skill against climatology 24% (14% to 32%) and against persistence 42% (36% to 48%). Month by month, skill against climatology ranged from 8% to 41%. On the 144 days that can be bias-corrected, the correction moves RMSE from 1.31 to 1.14 m/s.
AIFS 2, solar energy, day 3, over 174 days: skill against climatology 47% (41% to 53%) and against persistence 60% (56% to 63%). Month by month, skill against climatology ranged from 26% to 51%. On the 143 days that can be bias-corrected, the correction moves RMSE from 0.77 to 0.78 kWh/m²/day.
AIFS 1.1, 2 m temperature, day 3, over 248 days: skill against climatology 61% (56% to 65%) and against persistence 67% (64% to 69%). Month by month, skill against climatology ranged from 46% to 69%. On the 218 days that can be bias-corrected, the correction moves RMSE from 1.53 to 1.48 °C.
AIFS 1.1, 10 m wind speed, day 3, over 248 days: skill against climatology 29% (25% to 33%) and against persistence 47% (43% to 50%). Month by month, skill against climatology ranged from 8% to 37%. On the 218 days that can be bias-corrected, the correction moves RMSE from 1.41 to 1.27 m/s.
AIFS 1.1, solar energy, day 3, over 247 days: skill against climatology 45% (38% to 50%) and against persistence 59% (54% to 62%). Month by month, skill against climatology ranged from 22% to 52%. On the 216 days that can be bias-corrected, the correction moves RMSE from 0.58 to 0.59 kWh/m²/day.