Models
Model Comparison
Computer weather models, side by side for Faribault. The report card shows which ones have been closest to what actually happened at the Faribault airport.
The model comparison is unavailable right now. It refreshes every 30 minutes.
Average miss against the Faribault airport, °F
No model has been clearly closest on highs over the last 30 days: GFS has been off by 2.9° on average and NAM by 3.1°, a gap smaller than their day-to-day ups and downs.
No model has been clearly closest on lows over the last 30 days: NAM has been off by 3.6° on average and HRRR by 3.7°, a gap smaller than their day-to-day ups and downs.
No model has been clearly closest on highs over the last 90 days: NAM has been off by 2.4° on average and ICON by 2.5°, a gap smaller than their day-to-day ups and downs.
No model has been clearly closest on lows over the last 90 days: NAM has been off by 3.0° on average and HRRR by 3.1°, a gap smaller than their day-to-day ups and downs.
NBM has caught the most rain days at day 1 over the last 90 days: 21 of 26 (tied with Euro; a tie goes to fewer dry forecasts, then the smaller amount miss).
No model has been clearly closest on highs over the last year: ICON has been off by 2.3° on average and NBM by 2.4°, a gap smaller than their day-to-day ups and downs.
No model has been clearly closest on lows over the last year: NBM and NAM have both been off by 2.9° on average.
NBM has caught the most rain days at day 1 over the last year: 65 of 83.
ICON has caught the most days with 1 inch or more of snow at day 1 over the last year: 5 of 9 (tied with NBM; a tie goes to fewer false alarms, then the smaller amount miss).
Over the last 90 days, 7 of the 8 models ran too cold on highs at the Faribault airport, by 1.2° to 2.0° on average.
Over the last 90 days, 7 of the 8 models ran too warm on lows at the Faribault airport, by 2.0° to 4.5° on average.
Over the last year, 6 of the 8 models ran too cold on highs at the Faribault airport, by 1.0° to 1.5° on average.
“Runs warm” or “runs cold” by a name: over the last 90 days that model’s day-1 forecast leaned that way by 1° or more on average, steadily enough to stand out from its day-to-day misses. It shows only when the period shown leans the same way by 1° or more.
“Runs warm” or “runs cold” by a name: over the last 90 days that model’s day-1 forecast leaned that way by 1° or more on average, steadily enough to stand out from its day-to-day misses. It shows only when the period shown leans the same way by 1° or more.
Ranked by each model’s Lead day How many days ahead a forecast was made: day 1 is the forecast for tomorrow. This site's report card Average miss How far a forecast high or low has been from the Faribault airport's reading, on average, whether too warm or too cold. This site's report card over the period shown, fewest degrees first; a model needs 7 days scored to be ranked. The NWS isn’t ranked.
Ranked by rain days caught at day 1, most first; a tie goes to fewer dry forecasts, then the smaller amount miss. A model needs 7 rain days to be ranked.
Ranked by days with 1 inch or more caught at day 1, most first; a tie goes to fewer false alarms, then the smaller amount miss. A model needs 7 such days to be ranked.
Updated Nov. 2, 8:30 a.m.
Gray numbers come from fewer than 7 days: a snapshot, not a ranking. A model can miss badly one day and do fine the next.
Gray cells without a number: past what that row reaches. HRRR and NAM: archived runs reach day 1 only. ICON: its 180-hour run ends inside day 7. NWS: its 7-day forecast reaches day 6.
Gray cells without a number: past what that row reaches. HRRR: archived runs reach day 1 only.
Gray cells without a number: past what that row reaches. HRRR: archived runs reach day 1 only.
NWS, not ranked: scored since Sept. 29, 2026, from its 10:40 a.m. forecast each day (9:40 a.m. in winter); the NWS isn’t archived.
Rain scored against radar-plus-gauge analysis (NCEP Stage IV).
A rain day is 0.05 inch or more, midnight to midnight. Rain days caught: of the days it rained, how many the model forecast rain for. Dry when rain was forecast: of the days the model forecast rain, how many stayed dry. Amount miss: the average difference in inches on days either the model or the analysis had rain. The analysis is NCEP Stage IV (radar estimates corrected by rain gauges) at Faribault’s point, from the Iowa Environmental Mesonet; model numbers come from archived runs (Open-Meteo’s Previous Runs), days 1 to 3. NAM and Euro (9 km) have no rain in that archive.
Snow scored against the Faribault co-op station's 7 a.m. reports.
A day is 7 a.m. to 7 a.m., the observer's day. Snow days caught: of the days with 1 inch or more at the Faribault co-op station, how many the model forecast 1 inch or more for. False alarms: of the days the model forecast 1 inch or more, how many got less. Average miss and bias (forecast minus what fell, + = too snowy): inches, on days either the model or the observer had 0.1 inch or more. Our gauge: the average miss against CoCoRaHS MN-RC-45, which reports some mornings, not all. The model rows come from archived model runs (Open-Meteo’s Previous Runs, back to January 2024; the GFS’s snowfall from October 2024): for day 1, the snow each model forecast a day before each hour, as the model itself gives it. The Euro’s archive has no snowfall, so its snow is estimated from its rain and temperature (a snow ratio by temperature); it is shown last, marked, and never ranked with the others. The NWS row comes from its gridpoint forecasts saved by this site each morning since Oct. 7, 2026. Days the co-op didn’t report are left out, never counted as no snow. A window shows once it has 10 days with measurable snow.
Warm or Cold Habit
Day 1 average, forecast minus what happened
Leaned furthest: HRRR, 4.2° too cold.
Leaned furthest: Euro, 5.0° too warm.
Leaned furthest: GFS, 2.2° too warm.
Leaned furthest: Euro, 4.5° too warm.
Leaned furthest: NAM and GEM, 1.5° too cold.
Leaned furthest: ICON, 2.1° too warm.
How Often It’s Close
Day 1 forecasts within 2° and within 5° of the airport’s reading
Within 2° most often: GEM, 36% of days.
Within 2° most often: GFS and ICON, 39% of days each.
Within 2° most often: NBM, 48% of days.
Within 2° most often: HRRR, 42% of days.
Within 2° most often: ICON, 54% of days.
Within 2° most often: NBM, 46% of days.
Last 30 Days, Day by Day
Day 1 miss each day, 0° to 10°; dashed line at 2°; wins at right
Most wins: GEM, 7 of 28 days.
Most wins: NAM, 10 of 28 days.
A win is the closest day-1 forecast of the day; a tie goes to the 30-day leader. The NWS isn’t counted.
Report card is updating. Check back later.
The largest day-1 miss in the last 30 days, from the run this site had by the morning before
Euro’s low for Tuesday, Oct. 20. Its run from 1 a.m. Oct. 19 (06Z), which reached this site at 8:35 a.m., said 45.7°. The airport’s low that day was 32.5° (midnight to midnight standard time), so the forecast was 13.2° too warm.
For the week in plain English, with how sure each day is, see the 7-day forecast.
Each model’s newest run, in Central time, and when it reached this site
- HRRRMon. 11 p.m. run (Z time Coordinated Universal Time (UTC), also called Zulu: the local time at the prime meridian, which passes through Greenwich, England. National Weather Service glossary)Reached this site 1:05 a.m.
- NAMMon. 6 p.m. run (00Z)Reached this site Mon. 9:05 p.m.Tomorrow’s high 56°, 2° lower than the Mon. noon run
- GFSMon. 6 p.m. run (00Z)Reached this site Mon. 11:35 p.m.Tomorrow’s high 52°, same as the Mon. noon run
- Euro (9 km)Mon. 6 p.m. run (00Z)Reached this site 1:05 a.m.Tomorrow’s high 54°, 2° higher than the Mon. noon run
- EuroMon. 6 p.m. run (00Z)Reached this site 2:05 a.m.
- ICONMon. 6 p.m. run (00Z)Reached this site Mon. 10:05 p.m.Tomorrow’s high 56°, 1° lower than the Mon. noon run
- GEMRun time not publishedReached this site Mon. 1:35 p.m.
- NBMmidnight run (06Z)Reached this site 1:35 a.m.Tomorrow’s high 52°, same as the Mon. 11 p.m. run
- NWSIssued 12:26 a.m.
Next 5 days, from the numbers below
- Closest on Thursday: highs 47° to 51°.
- Widest split on Saturday, 14° apart: 42° (GEM) to 56° (GFS).
- Rain through Friday: from 0.58 inches (NBM) to 0.90 inches (GFS).
°F, each model as far as it goes
Models’ middle high tomorrow: 55° (6 models 52° to 56°). Our forecast ›
Each model’s last 8 runs, °F
Biggest move: GFS, from 59° to 52° over its last 8 runs.
Biggest move: ICON, from 48° to 50° over its last 8 runs.
Biggest move: ICON, from 55° to 51° over its last 8 runs.
Biggest move: GFS, from 47° to 44° over its last 8 runs.
Biggest move: GFS, from 55° to 49° over its last 8 runs.
Biggest move: Euro (9 km), from 40° to 33° over its last 8 runs.
Biggest move: ICON, from 45° to 49° over its last 8 runs.
Biggest move: Euro (9 km), from 40° to 32° over its last 8 runs.
Biggest move: Euro (9 km), from 47° to 57° over its last 8 runs.
Biggest move: Euro (9 km), from 41° to 40° over its last 8 runs.
Each dot is one run, placed at the time the run started. Days run midnight to midnight Central Standard Time, as in the Report Card. GEM is left out because its run times are not published.
Next 7 days, each model’s newest run
Highest dew point: 63°, today at 2 a.m. (GEM).
Least cloud: Sunday, 16% on average across the models.
Strongest gust: 42 mph, Sunday at noon (Euro).
Tap or drag across a chart (or hover) to read the models’ middle and range for that hour.
Daily total, inches, next 5 days
Wettest day: Wednesday, 0.17 in. (NWS) to 0.89 in. (GFS).
Daily peak gust, mph, next 5 days
Strongest gusts: today, 19 mph (NBM) to 37 mph (NAM).
Daily snowfall, inches, next 5 days
No snow of 0.1 in. or more in any model through Saturday.
Model estimate. Snow-to-water ratios vary.
The shakiest call in the next 7 days: Friday night’s low. Not sure: GFS, GEM, NBM and ICON 10° apart. On the 7-day forecast ›
Daily high, every ensemble run, °F
Chances from the ensemble runs
Each dot is one run of an Ensemble A set of runs of one model, started from slightly different conditions, that shows the range of possible outcomes. National Weather Service glossary. The bar under the dots spans the middle 80% of its runs (10th to 90th percentile), and the tick is the middle run.
Each thin line is one run of an ensemble. The thick lines are each ensemble’s average; the Euro’s is dashed.
The air above Faribault, next 48 hours
Across the Sounding A set of measurements of the air from the ground up, such as temperature, humidity, pressure and wind, at one time. National Weather Service glossary for the next 48 hours, the lowest Lifted index A measure of instability. Negative values mean unstable air: the more negative, the stronger any storms' updrafts are likely to be. National Weather Service glossary is 0, Wednesday at 3 a.m. (HRRR).
Words in the table: CAPE Convective available potential energy: the energy available to rising air in storms. Higher values mean greater potential for severe weather. National Weather Service glossary · CIN Convective inhibition: the energy needed to start air rising into storms, a measure of how strong the lid on them is. National Weather Service glossary · Wind shear The change in the wind's direction and speed with height, a key factor in whether severe thunderstorms develop. National Weather Service glossary · Precipitable water The depth of water at the ground if all the water vapor in the air above fell out, usually counted up to 300 mb. National Weather Service glossary · Freezing level The height at which the air temperature first drops below freezing. National Weather Service glossary · Millibar (mb) A unit of air pressure: one thousandth of a bar. National Weather Service glossary
Next 48 hours, each model’s newest run
Most instability: 100 J/kg, Wednesday at midnight (Euro).
Most precipitable water: 1.41 in., today at 9 a.m. (HRRR).
Lowest freezing level: 7,700 ft above ground, Wednesday at 3 p.m. (GFS).
Tap or drag across a chart (or hover) to read each model’s value for that time.
Climate Prediction Center outlooks for Faribault
Climate outlook The Climate Prediction Center's chances that conditions over a period will be below normal, near normal or above normal. National Weather Service glossary for Nov. 10 to Nov. 16: temperature above normal (70% chance), precipitation near normal (36% chance).
Three daily weather patterns: last 30 days and the GFS ensemble forecast
Latest (Oct. 31): Arctic Oscillation (AO) Air pressure over the Arctic and the middle latitudes swinging between phases; the negative phase lets cold air plunge into the Midwest. National Weather Service glossary +0.8, North Atlantic Oscillation (NAO) A large swing in air pressure between the high near the Azores and the low near Iceland, which steers storms across the North Atlantic. National Weather Service glossary +1.1, Pacific-North American pattern (PNA) A pattern of high and low pressure aloft over the Pacific and North America. Its positive phase warms western Canada and cools the Southeast. NOAA Climate Prediction Center +0.6. The GFS ensemble’s average for Nov. 17: AO −0.2, NAO −0.1, PNA +0.4.
Storm Prediction Center, next 3 days
Convective outlook The Storm Prediction Center's forecast of where thunderstorms are expected and how severe they may be, issued several times a day. National Weather Service glossary, next 3 days: no severe risk for Faribault; general thunderstorms today.
What each model is, from the pages that state it
| Model | Covers | Grid spacing bar: 0 to 25 km | Runs | Forecast length bar: 16 days, first 7 shaded | Time steps |
|---|---|---|---|---|---|
| HRRRNOAA | Contiguous U.S. | 3 km | Every hour | 18 hours (48 hours for the 0Z, 6Z, 12Z and 18Z runs) | Hourly |
| NAMNOAA | Contiguous U.S. | 3 km | Every 6 hours | 60 hours | Hourly |
| GFSNOAA | Global | About 13 km | Every 6 hours | 16 days | Hourly, 3-hourly after 120 hours |
| Euro (9 km)ECMWF | Global | 9 km | Every 6 hours | 15 days | Hourly, 3-hourly after 90 hours, 6-hourly after 144 hours |
| EuroECMWF | Global | About 25 km | Every 6 hours | 15 days | 3-hourly, 6-hourly after 144 hours |
| ICONDWD, Germany’s weather service | Global | About 11 km | Every 6 hours | 7½ days | Hourly, 3-hourly after 78 hours |
| GEMCanada’s weather service | Global | About 15 km | Every 12 hours | 10 days | 3-hourly |
| National Blend of Models (NBM) Forecast guidance from the NWS, calibrated and blended from many weather models, the NWS's own and others'. National Weather Service glossaryNOAA | Contiguous U.S. | 2.5 km | Every hour | 11 days | Hourly, 3-hourly after 36 hours, 6-hourly after 8 days |
| NWSTwin Cities, MN office (Chanhassen) | 7 days | Day and night | |||
| Ensembles | |||||
| GFS ensembleNOAA, 31 runs | Global | About 25 km | Every 6 hours | 10 days | 3-hourly |
| Euro ensembleECMWF, 51 runs | Global | About 25 km | Every 6 hours | 15 days | 3-hourly |
Two Euros: “Euro (9 km)” is the one the temperature chart, rain, wind and snow on this page use; “Euro” is the coarser one the report card also scores. How sure uses the GFS and Euro ensembles. Sources, read Oct. 20, 2026: Open-Meteo, NOAA models; Open-Meteo, ECMWF; Open-Meteo, DWD ICON; Open-Meteo, GEM; Open-Meteo, ensembles; the NWS API's office record. A cell left empty is one its source doesn’t state.