Winter 2026-2027. Historic Potential

[Weather Will]

Every couple of weeks will post the latest WB CFSV2 30 day monthly anomaly maps.

[Stormchaserchuck1]

I wish lol. You have to wonder why these models are so bad

[Stormchaserchuck1]

9 hours ago, psuhoffman said:

BWI's average high on Xmas since 2015 is 52.  December has been the month impacted the most by warming.  The fact that SST's have warmed impacts the start of winter due to lag in water temps cooling.  This has been documented...someone published monthly snowfall changes over the last 50 years and our area was the worst shade for December.  January to March has been less impacted.  Most of our decrease in snowfall climo has come from losing early in the season.  It is what it is.  

February has by far been the worst if you narrow it down to since 2018. 

For some reason there has been incredible -PNA tendency in Feb and March for the last 9 years. Not to mention 13 straight +NAO Feb's, with 10/13 being >1.00! Take the indexes out and I agree December has been the GW month

[GaWx]

2 hours ago, Weather Will said:

Every couple of weeks will post the latest WB CFSV2 30 day monthly anomaly maps.

WB CFSv2 temp maps:

1. Have algo issues that lead to bad artifacts. I’ve posted on this here in the past. Here are some examples

-note the cold spot over S Lake Michigan and warm spot only 250 miles to the north over N Lake Michigan that are almost always on these WB maps; the real CFSv2 output essentially never has these

-note the E coast down to FL usually looks like a thin sharp warm barrier to any colder anomalies to the west; again, this essentially never is on the real output

2. WB CFSv2 maps are strongly cold biased for much of the E US. And this is the case even with WB using the colder 1981-2010 climo instead of 1991-2010, which the CFSv2 site uses! All one has to do to know this is to compare to the maps at the CFSv2 site.
 

 All 4 (DJFM) are notably too cold in almost all of the area away from the E coast. But I’ll focus on the Dec map to illustrate the strong cold bias:

-Chicago’s +0.5C/1F vs 1981-2010 on WB but the CFS site’s relevant 10 day mean maps (see below) are all significantly warmer and they average +2.25C/4F vs the warmer 1991-2020 meaning WB is ~4F too cold.

-SL’s 0.0 C/F vs 1981-2010 on WB but CFS actual maps are all significantly warmer and average +1.5C/3F vs 1991-2020 meaning WB is ~4F too cold

-Knoxville’s -1C/-2F vs 1981-2010 on WB but CFS actual maps have it at +0.6C/+1F vs 1991-2020 meaning WB’s ~4F too cold

-Despite the CFS site’s maps averaging AN from TN-NC north, WB in inland areas doesn’t have AN start til 500 miles to the N in N IL to N OH to NW PA and that’s even vs a colder climo base! That’s ignoring the immediate E coast, which has its own problems as already noted and is actually a bit too warm overall despite it being 4F too cold for much of the inland areas!

Relevant 10 day mean Dec maps at CFSv2 site:

9/10-19:

 

8/31-9/9:

 

8/21-30:

 

8/12-21:

@Stormchaserchuck1@donsutherland1

[Stormchaserchuck1]

A big issue I see with the CFS is, with the Winter 500mb pattern being modeled over the N. Pacific, you aren't going to get that strong of an East Coast trough. Since 2016 the Pacific has held much more weight than the NAO. Then again as I already mentioned in this thread, in 2024 and 2025 I saw runs of the CFS having -38F anomalies in the NW, US for the whole month of December! Can you believe they ended up well above average? 

[bncho]

5 hours ago, mitchnick said:

On a related note, a met from the NE forum put together a snow storm archive of winters going back to 92/93. It is NE centric, but plenty of radars and maps that include the Conus and MA region. Here's a link if interested, assuming it hasn't already been posted.

https://www.jdjweatherconsulting.com/storm-archive

This winter, all it will take is one storm.

[Stormchaserchuck1]

"All it takes is one storm" So you are stalking out Winter for 6 months for 18 hours of snowfall? No thanks, give me 20 light snow events, "snowfall days". My trees are already losing leaves - the most depressing 6 months of the year for 1 storm, no lol. 

[dailylurker]

36 minutes ago, Stormchaserchuck1 said:

"All it takes is one storm" So you are stalking out Winter for 6 months for 18 hours of snowfall? No thanks, give me 20 light snow events, "snowfall days". My trees are already losing leaves - the most depressing 6 months of the year for 1 storm, no lol. 

You must be new here lol. 

[Eskimo Joe]

2 hours ago, Stormchaserchuck1 said:

"All it takes is one storm" So you are stalking out Winter for 6 months for 18 hours of snowfall? No thanks, give me 20 light snow events, "snowfall days". My trees are already losing leaves - the most depressing 6 months of the year for 1 storm, no lol. 

That's how it's worked since 2016 bro. New normal.

[Ji]

BWI's average high on Xmas since 2015 is 52.  December has been the month impacted the most by warming.  The fact that SST's have warmed impacts the start of winter due to lag in water temps cooling.  This has been documented...someone published monthly snowfall changes over the last 50 years and our area was the worst shade for December.  January to March has been less impacted.  Most of our decrease in snowfall climo has come from losing early in the season.  It is what it is.  

Look at the February stats. They’ve been as bad as December.

[Bob Chill]

3 hours ago, Ji said:

Look at the February stats. They’ve been as bad as December.

So we're (over)due then. Boom!

[WxUSAF]

29 minutes ago, Bob Chill said:

So we're (over)due then. Boom!

Very much down with a Super Bowl weekend KU

[donsutherland1]

14 hours ago, GaWx said:

WB CFSv2 temp maps:

1. Have algo issues that lead to bad artifacts. I’ve posted on this here in the past. Here are some examples

-note the cold spot over S Lake Michigan and warm spot only 250 miles to the north over N Lake Michigan that are almost always on these WB maps; the real CFSv2 output essentially never has these

-note the E coast down to FL usually looks like a thin sharp warm barrier to any colder anomalies to the west; again, this essentially never is on the real output

2. WB CFSv2 maps are strongly cold biased for much of the E US. And this is the case even with WB using the colder 1981-2010 climo instead of 1991-2010, which the CFSv2 site uses! All one has to do to know this is to compare to the maps at the CFSv2 site.
 

 All 4 (DJFM) are notably too cold in almost all of the area away from the E coast. But I’ll focus on the Dec map to illustrate the strong cold bias:

-Chicago’s +0.5C/1F vs 1981-2010 on WB but the CFS site’s relevant 10 day mean maps (see below) are all significantly warmer and they average +2.25C/4F vs the warmer 1991-2020 meaning WB is ~4F too cold.

-SL’s 0.0 C/F vs 1981-2010 on WB but CFS actual maps are all significantly warmer and average +1.5C/3F vs 1991-2020 meaning WB is ~4F too cold

-Knoxville’s -1C/-2F vs 1981-2010 on WB but CFS actual maps have it at +0.6C/+1F vs 1991-2020 meaning WB’s ~4F too cold

-Despite the CFS site’s maps averaging AN from TN-NC north, WB in inland areas doesn’t have AN start til 500 miles to the N in N IL to N OH to NW PA and that’s even vs a colder climo base! That’s ignoring the immediate E coast, which has its own problems as already noted and is actually a bit too warm overall despite it being 4F too cold for much of the inland areas!

Relevant 10 day mean Dec maps at CFSv2 site:

9/10-19:

 

8/31-9/9:

 

8/21-30:

 

8/12-21:

@Stormchaserchuck1@donsutherland1

It's unfortunate. WB will never correct the algorithms. 

[Eskimo Joe]

Always a great time when the ultimate snow weenie has thrown in the towel in September 

 

[Bob Chill]

2 hours ago, Eskimo Joe said:

Always a great time when the ultimate snow weenie has thrown in the towel in September 

 

Can't remember how long the Siberian snow thing has been around but its been at least 12-14 years now. Iirc, Cohen has busted on winter forecasts a good bit more than got it right. Based on that statistic, his latest post is actually a net positive for winter snowfall chances. 

[Steve25]

10 hours ago, Stormchaserchuck1 said:

"All it takes is one storm" So you are stalking out Winter for 6 months for 18 hours of snowfall? No thanks, give me 20 light snow events, "snowfall days". My trees are already losing leaves - the most depressing 6 months of the year for 1 storm, no lol. 

To answer your question…absof******lutely! 
 

Considering I haven’t experienced a snowstorm over 6 inches since 2016. I’ve seen everything else since 2016. Brutal cold stretches leading to frozen bodies of water, frequent light snow events, a moderate snow event that sticks around for weeks due to the cold that followed. I need the big dog. I would  easily sacrifice 90% of the winter to get one MECS/HECS. I need that feeling again. It’s been so long. 
 

I will say that I come into the season with a different perspective than you overall, though. I don’t consider this to be the most depressing 6 months of the year. I’d take October-March over April-September pretty much every time, no matter what the specific winter weather circumstances are. 

[osfan24]

12 hours ago, Stormchaserchuck1 said:

"All it takes is one storm" So you are stalking out Winter for 6 months for 18 hours of snowfall? No thanks, give me 20 light snow events, "snowfall days". My trees are already losing leaves - the most depressing 6 months of the year for 1 storm, no lol. 

Good luck with 20 light snow events in a Nino this strong. This is a boom or bust winter, and a matter of how many booms we could get if it does boom.

[Terpeast]

In a super nino year, if I could get one big dog that beats my PR of 30”, the rest of the winter can be in the 60s for all I care. 

[CAPE]

21 hours ago, GaWx said:

WB CFSv2 temp maps:

1. Have algo issues that lead to bad artifacts. I’ve posted on this here in the past. Here are some examples

-note the cold spot over S Lake Michigan and warm spot only 250 miles to the north over N Lake Michigan that are almost always on these WB maps; the real CFSv2 output essentially never has these

-note the E coast down to FL usually looks like a thin sharp warm barrier to any colder anomalies to the west; again, this essentially never is on the real output

2. WB CFSv2 maps are strongly cold biased for much of the E US. And this is the case even with WB using the colder 1981-2010 climo instead of 1991-2010, which the CFSv2 site uses! All one has to do to know this is to compare to the maps at the CFSv2 site.
 

 All 4 (DJFM) are notably too cold in almost all of the area away from the E coast. But I’ll focus on the Dec map to illustrate the strong cold bias:

-Chicago’s +0.5C/1F vs 1981-2010 on WB but the CFS site’s relevant 10 day mean maps (see below) are all significantly warmer and they average +2.25C/4F vs the warmer 1991-2020 meaning WB is ~4F too cold.

-SL’s 0.0 C/F vs 1981-2010 on WB but CFS actual maps are all significantly warmer and average +1.5C/3F vs 1991-2020 meaning WB is ~4F too cold

-Knoxville’s -1C/-2F vs 1981-2010 on WB but CFS actual maps have it at +0.6C/+1F vs 1991-2020 meaning WB’s ~4F too cold

-Despite the CFS site’s maps averaging AN from TN-NC north, WB in inland areas doesn’t have AN start til 500 miles to the N in N IL to N OH to NW PA and that’s even vs a colder climo base! That’s ignoring the immediate E coast, which has its own problems as already noted and is actually a bit too warm overall despite it being 4F too cold for much of the inland areas!

Relevant 10 day mean Dec maps at CFSv2 site:

9/10-19:

 

8/31-9/9:

 

8/21-30:

 

8/12-21:

@Stormchaserchuck1@donsutherland1

He keeps posting them even though it has been said over and over that WB sucks ass wrt the CFS depiction, esp at the surface.

[Terpeast]

3 minutes ago, CAPE said:

He keeps posting them even though it has been said over and over that WB sucks ass wrt the CFS depiction, esp at the surface.

I don’t even check the surface for cfs maps, just 500mb and draw my own conclusions 

[CAPE]

3 hours ago, Bob Chill said:

Can't remember how long the Siberian snow thing has been around but its been at least 12-14 years now. Iirc, Cohen has busted on winter forecasts a good bit more than got it right. Based on that statistic, his latest post is actually a net positive for winter snowfall chances. 

He is awful. Been obvious ever since he took credit for that insane month(Feb?) several winters ago when Boston got dumped on over and over and had like 100". None of his shit worked out that winter. Snake oil salesman.

[CAPE]

1 minute ago, Terpeast said:

I don’t even check the surface for cfs maps, just 500mb and draw my own conclusions 

Same. h5 is all I look at with these seasonal models. That said, this is a WB issue. Sucks at h5 too. Check it against TT.

[Terpeast]

12 minutes ago, CAPE said:

Same. h5 is all I look at with these seasonal models. That said, this is a WB issue. Sucks at h5 too. Check it against TT.

Haha i actually use TT most of the time it’s most mobile friendly. I like pivotal for desktop analysis but its not as mobile friendly

[CAPE]

Google AI-

In seasonal climate modeling, the Climate Forecast System (CFS) seasonal control run—a single, unperturbed deterministic simulation extended out to nine months—is highly prone to systematic model biases, underdispersion, and severe run-to-run volatility. While the control run provides a baseline, its raw output is widely considered unreliable for direct long-range forecasting without extensive statistical bias correction and ensemble averaging.

WW doesnt understand control runs. But even the CFS monthly(M) that goes out 3 months is off on WB.

[CAPE]

That's similar to using a single GFS op run and extending it out a month or more. We know how bad it sucks at 10 days lol.

[MillvilleWx]

Just now, CAPE said:

That's similar to using a single GFS op run and extending it out a month or more. We know how bad is sucks at 1 day lol.

FYP

[CAPE]

12 hours ago, Eskimo Joe said:

That's how it's worked since 2016 bro. New normal.

La Nina ftw. At least over here, and esp right at the coast.