Skip to main content

No, there will not be a September snowstorm in Minnesota

The American weather model has lost its mind and it cannot be trusted this far out.

Before you see something making the rounds on social media that might make you spit out your coffee, I’m going to explain why it’s complete garbage and why anyone posting it with any seriousness is engaging in bad science. I’m talking about the American (GFS) model’s snowfall output for the end of September. 

The 12z run from the American model shows a snowstorm over Minnesota on Sept. 28. It is garbage. 

The 12z run from the American model shows a snowstorm over Minnesota on Sept. 28. It is garbage. 

This of course might get lots of clicks and shares. After all, we do love to complain about our climate in Minnesota. Luckily, we have almost a century and a half of climate data to put everything into perspective, in addition to a pile of much better weather forecasting computer models. And it's early enough in the season where most sane people won’t take it seriously. Here's why you need to dismiss it.


This model is for 15 days out

Any forecast showing anything in detail past 7-10 days is mere entertainment in the meteorological world. So, even in January, when snow is a reality, anyone posting any snow forecast that’s more than a week or 10 days out is usually very low on the confidence scale. The models become exponentially less accurate with time because all the little errors that go into the model get multiplied over and over with time.


The American model has been terrible

The American (GFS) model has stunk – really bad – lately. If you’re a casual weather junkie and look at the models occasionally, you’ve probably noticed it forecasting triple-digit high temperatures multiple times in its 16-day forecast every day since June 1. 

We only hit 100+ once this summer, so it doesn’t have a great track record. The same model run from that produces the snowstorm in late September also has the TWin Cities hitting 103 degrees on Sept. 20 amid a string of 90s.

The American model's temperature forecast from the 12z run, which is also likely garbage. 

The American model's temperature forecast from the 12z run, which is also likely garbage. 

While it might be kind of exciting to go from 103 to a snowstorm in a week, it’s statistically very unrealistic. Darn near impossible, in fact. Here’s a look at it’s 24-hour high temperature Sept. 28 from the same update that produced the 103-degree reading a week earlier: 37 in the Twin Cities on Sept. 28. That's extreme – and extremely unlikely to happen. 

gfs-deterministic-ncentus-t2m_f_max_last24-4366400-2

Here’s the EXACT same model but the latest update, a mere six hours later: So yes, it went from producing a Sept. 28 high in the Twin Cities of 37 degrees to 88. You just can't take it seriously. 

gfs-deterministic-ncentus-t2m_f_max_last24-4366400

A future article is probably a good idea to explain the strengths and weaknesses of all the models, but the American (GFS) model has really fallen behind. The European model is a much more accurate model, partly because it has more computing power jammed into 10 days versus the lesser computing power spread out over 16 days in the American (GFS) model.


Model consensus shows no snowstorm

Consensus! One of the things we do in a forecast is look at any consensus in the models. I talk about the ensemble models occasionally, and those are the multiple simulations of the same model with slightly different conditions, from which probabilities are derived. In other words, if we tweak the model and run it a bunch of times, what happens? What’s the consensus? We can do this with each of the three marquee medium range models (European, Canadian, American).

Let’s compare. I’ve outlined in red how the American model compares to the other medium range models for Sept. 27-29. The American model is clearly an outlier. 

7195B2D6-5228-4876-994D-8D946B18F6AC

Climatology says a snowstorm won't happen

Regardless of what any model might say, we always have statistics. It has snowed in late September in the Twin Cities before, but only three times in 137 years of record keeping. And two of the three snowfalls were under a half inch, while the other was just 1.7 inches. 

The average first measurable (0.1”+) snowfall in the Twin Cities is Nov. 7. 

SNOW probs

You can see that the chance of the first measurable snow is exponentially more likely as we head into late October and early November. By Nov. 18,  75% of the time we’ve had our first snowfall of at least 0.1 inches. 

It is ten times more likely we have the first snowfall by Nov. 7 than by Oct. 12. Clearly, no matter the year, there’s a big shift that happens in Minnesota weather in mid-October.

Of course, you don’t have take my word for it. Other meteorologists are having fun with it, too, including National Weather Service meteorologist Melissa Dye. 

And a student of meteorology, Chris Wicklund. 

The snow will come, my friends, but not in the next two weeks. It’s far more likely that we end up with a September that goes down as one of the ten warmest in recorded history (it’s well on track at this point already). But take these lessons above to heart with any dramatic forecast you see in the future. 

Next Up

Screen Shot 2022-09-28 at 3.36.57 PM

Scott County attorney candidates differ on marijuana possession

Scott County is one of a few Minnesota counties where the incumbent prosecutor is facing a challenger in November.

Screen Shot 2022-09-23 at 10.21.38 PM

Teen charged in shooting outside of Richfield football game

Due to the age of the suspect, the case won't be made public.

Screen Shot 2020-02-06 at 9.05.08 AM

Twin Cities Summer Jam no more: popular festival brought to an end

The festival noted how the land they host camping on has recently been sold.

image

Expect major traffic delays on Renaissance Festival's final weekend

Scott County officials said road closures will be to blame.

Voting survey vote election

Ramsey County sends out ballots with deceased candidate on it

The county said it's addressing the issue with the Minnesota Supreme Court.

Screen Shot 2022-09-28 at 12.30.28 PM

New haunted house experience coming to Inver Grove Heights

The exhibit will feature family-friendly and 18+ events.

Wayzata High School

Niche rankings identify best Minnesota schools, districts

Eden Prairie has schools ranked in both the private and public categories.

unnamed

Walker, The Current announce end of Rock the Garden festival

The Walker Art Center and Minnesota Public Radio announced the festival will retire so the organizations can dream up new events.

Related

Screen Shot 2022-03-22 at 8.15.20 AM

Storm to dump rain, snow, freezing rain on Minnesota

Wintry weather will cause the most significant impacts in the Northland.

storm

Storms could go severe Saturday and Sunday in Minnesota

Severe storms are possible Saturday night and again Sunday afternoon/night in Minnesota.

Screen Shot 2022-03-18 at 12.44.42 PM

Rain and snow looks likely next week in Minnesota

Meteorologist Sven Sundgaard shows what 8 different models are suggesting in terms of snowfall.

Minneapolis skyline

Climate change in Minnesota is clearly seen at night

A lot of warm days and far fewer chilly nights.

822 Thumbnail (1)

Atmospheric waves could bring storms to MN this week

The stronger of the two waves is expected to arrive this weekend.

cold, wind chill, freezing

How a rare 'triple dip' La Nina could impact Minnesota's winter

Three consecutive La Nina winters has only happened twice since 1950.

Lake Nokomis

Why Minnesota's cool spring is a good indicator for a hot summer

Meteorologist Sven Sundgaard explains why our cool spring means there's a good chance Minnesota has a hot summer.