Pattern January 2020 - Operation Thaw Alaska

[KyloG]

Maybe they'll be right. However, I'm starting to develop an opinion that while analogues provide some value, it might be less than we think. I've seen lots of well-reasoned analogs show why a favorable pattern should move into place that have failed to work out. It seems like only analogs that show warm have a way of working out. I wonder why that is? Maybe it's less about them being good analogs and more about the fact that we're just tilted to warm now.

I don't know how useful analogs really are anymore. I think if you just lean warm, you're going to end up mostly right, which makes it look like the warm analogs you rolled out are right on the money. Now, after reading all of that, it seems somewhat paradoxical, maybe even nonsensical. But I'll post it anyway and let the chips fall.

Any analog pre-1990 seems rather low probability. I can't think of a winter in the past 20 years that we said...oh this is like a 1960's winter.

 

[Rain Cold]

Any analog pre-1990 seems rather low probability. I can't think of a winter in the past 20 years that we said...oh this is like a 1960's winter.

I'm not even sure analogs before 2010 are very good anymore. Either way, you're left with a very low sample size to choose from, which ends up skewing your result set.

 

[KyloG]

I'm not even sure analogs before 2010 are very good anymore. Either way, you're left with a very low sample size to choose from, which ends up skewing your result set.

I think you are right. Including this winter.... 7/9 past winters have had a GOA/Aleutian ridge driven pattern. We were fortunate in 2014 the ridge was more poleward, but every other winter has been a vomit fest. 2015 was a nice change, weak nino pattern, and 2016 the super duper nino overwhelmed everything. We really shouldn't be shocked at what's going on this winter...it's the same thing year after year.

But, it's a new decade...I think Feb/March will show us where we are headed for the next 10 years...

 

[Jon]

Let’s hope this thinking isn’t going to verify for the last half of the month:

“The 11-15 Day pattern discussed above may have some staying power
into the latter part of January from a historical perspective. The maps on
right shows 500H and temperature anomalies for analogs from CPC based on the GFS Ensemble Day 11 pushed forward for the 16-30 Day period. These analogs suggest a continued strong –PNA and +EPO signal, implying a continuation of above normal temperatures across the eastern half and belows in the Southwest.” from Maxar

Any thoughts about this? I know many will want to dismiss this simply because it goes against what they want to happen. And had this said the opposite, those same folks would be in total agreement and maybe even give a “like” to it.

Not dismissing it bc I don’t like it, this just doesn’t work.

A year or two back there was a guy on Twitter who auto generated those “monthly” long range CPC analogs based on the composite years. He had a website up and running and everything. Out to week 3 and week 4, maybe even week 5-6.

It’s a good idea, and I used to do it too, but because of the relative different individual correlations of said analog years and not to mention different background states, they didn’t have any accuracy what so ever. In fact the guy got rid of his website they were so bad and I stopped caring to do it. It’s fun and a neat idea, but it has no use in LR forecasting that I’m aware of.


Sent from my iPhone using Tapatalk

 

[Stormdoc]

GFS is trying to keep me paying attention to this trough, it keeps pulling me back in!

 

[Jon]

Not dismissing it bc I don’t like it, this just doesn’t work.

A year or two back there was a guy on Twitter who auto generated those “monthly” long range CPC analogs based on the composite years. He had a website up and running and everything. Out to week 3 and week 4, maybe even week 5-6.

It’s a good idea, and I used to do it too, but because of the relative different individual correlations of said analog years and not to mention different background states, they didn’t have any accuracy what so ever. In fact the guy got rid of his website they were so bad and I stopped caring to do it. It’s fun and a neat idea, but it has no use in LR forecasting that I’m aware of.


Sent from my iPhone using Tapatalk

So just to try it. I believe this is what Maxar was doing that @GaWx mentioned but maybe he can clarify. (oddly enough my results are nowhere close to what they were mentioning?)

This is taking the Date centered on the CPC 8-14 day analog (1/12/19) and using the 9 analog dates in the chart to generate a +20 day composite for 2/1/19...

Behold.




Going forward in time only TWO days, gives me this (went forward 2 days in each composite year, making this the supposed forecast for 2/3/2020 based on CPC's analog year composite)


2/4/2020


2/5/2020



2/6/2020




you get the idea. If we were to make a 4 day composite of those last 4 images, we would have quite possibly the coldest best pattern ever. So are we to expect that Feb 3-6? ...common sense would be screaming no.

So I used to do this and provide animations of the predicted pattern, etc. 9 times out of 10 some wild blockbuster pattern showed up and I thought I was onto the next generation of weather forecasting...problem was these composites never materialized.

We'll compare these next month and see how well they did. The issue is CPC 8-14 day analogs will create a centered date of 01/13/19 today and the composite years will be changed, which will change these maps and thus another solution for a period only 1 day different in lead time...

 

[FallsLake]

You can't say there's no cold air in the northern hemisphere. Hopefully it stays long enough until the pattern flips. If so, we'll have a quicker turn to cold in the SE.

 

[Jon]

You can't say there's no cold air in the northern hemisphere. Hopefully it stays long enough until the pattern flips. If so, we'll have a quicker turn to cold in the SE. View attachment 29509

This is what I don’t get with the complaints of the PV on the other side of the hemisphere by sites like Bamwx, we have cold on our side — just nothing to push it south!


Sent from my iPhone using Tapatalk

 

[KyloG]

Nice to see the GEFS split the trop PV into two pieces. #positivity

 

[SD]

Nice to see the GEFS split the trop PV into two pieces. #positivity


View attachment 29510

Gefs really wants to use the Pacific ridge to split the tpv then go over the pole and bridge with scandy riding. Seems like the model has a habit of doing this in the extended. Don't get me wrong its nice to see but until the long range eps/geps agree its the gefs doing gefs things. But it also shows you how you get out of the escape room

Sent from my SM-G975U using Tapatalk

 

[Chazwin]

This is what I don’t get with the complaints of the PV on the other side of the hemisphere by sites like Bamwx, we have cold on our side — just nothing to push it south!


Sent from my iPhone using Tapatalk

This is the main reason that I have remained mildly optimistic about the prospect of at least getting a 2-3 week window that we could see a major storm. The last several years there just simply had not been any true arctic air in our source regions to tap into, so that even when indicies got favorable or any mild hints of blocking showed up, there was no cold air available to take advantage of it.

 

[SD]

It'll be interesting to see how temps verify in the d8-14 range versus the ensemble mean temps. Taking the eps mean at face value it's a +8-10 torch but highs short of 60 aren't excessively warm in my opinion. If the warmer pattern does verify you'd have to think these temps are 5-10 too cool and being weighed down by members showing a cooler pattern. That leads me to the question of is there really any value in even looking at the mean temp past maybe D7

Sent from my SM-G975U using Tapatalk

 

[GaWx]

So just to try it. I believe this is what Maxar was doing that @GaWx mentioned but maybe he can clarify. (oddly enough my results are nowhere close to what they were mentioning?)

This is taking the Date centered on the CPC 8-14 day analog (1/12/19) and using the 9 analog dates in the chart to generate a +20 day composite for 2/1/19...

Behold.

View attachment 29503


Going forward in time only TWO days, gives me this (went forward 2 days in each composite year, making this the supposed forecast for 2/3/2020 based on CPC's analog year composite)
View attachment 29504

2/4/2020
View attachment 29505

2/5/2020
View attachment 29506


2/6/2020
View attachment 29507



you get the idea. If we were to make a 4 day composite of those last 4 images, we would have quite possibly the coldest best pattern ever. So are we to expect that Feb 3-6? ...common sense would be screaming no.

So I used to do this and provide animations of the predicted pattern, etc. 9 times out of 10 some wild blockbuster pattern showed up and I thought I was onto the next generation of weather forecasting...problem was these composites never materialized.

We'll compare these next month and see how well they did. The issue is CPC 8-14 day analogs will create a centered date of 01/13/19 today and the composite years will be changed, which will change these maps and thus another solution for a period only 1 day different in lead time...

Jon,
Nice analysis! OK, here are the Maxar maps for the CPC analogs, 2 meter temps and H5:

 

[Jon]

Jon,
Nice analysis! OK, here are the Maxar maps for the CPC analogs, 2 meter temps and H5:

View attachment 29511

View attachment 29512

Oh ok, interesting to see a -NAO on those maps between day 16-30. Maybe this will be exactly how it works out because you can infer we will be in a pattern change in that period. This is a great case study and maybe we can throw data around and look at the CPC analog archive and see if pattern changes and flips were correctly forecasted with this method. I’m going based solely on memory and I don’t think it was that great...here’s to hoping It works out this time


Sent from my iPhone using Tapatalk

 

[PisgahNCWeather]

Like the positivity fellas! Remember words do have power and it’s best to keep that in mind always.

 

[deltadog03]

Euro op so far looks like it might be different a bit. Starting to see some ridge start to build over the pole. And the ridge looks closer to the west coast a bit.

 

[deltadog03]

We go from this to HR 194 (so far this run)

 

[Webberweather53]

We go from this to HR 194 (so far this run)

View attachment 29515View attachment 29516

Yeah pieces of the building ridge over Scandinavia are getting sent over the pole in the long-range, if we can finally shove this NE Pac ridge into Alaska, we're in great shape to see a legit PV split at the tail end of Jan, with the stronger, primary lobe of the TPV on our side of the pond.

 

[deltadog03]

Yeah pieces of the building ridge over Scandinavia are getting sent over the pole in the long-range, if we can finally shove this NE Pac ridge into Alaska, we're in great shape to see a legit PV split at the tail end of Jan, with the stronger, primary lobe of the TPV on our side of the pond.

I agree. I know it’s only 1 run but the euro op really wanted to make this happen. We have all the cold on our side of globe. That’s a good thing to start with.

 

[Jimmy Hypocracy]

Yeah pieces of the building ridge over Scandinavia are getting sent over the pole in the long-range, if we can finally shove this NE Pac ridge into Alaska, we're in great shape to see a legit PV split at the tail end of Jan, with the stronger, primary lobe of the TPV on our side of the pond.

With that EPO doing its thing, If we were able to score a -AO would it only help to strengthen the -PNA or would ridging at the pole begin to drive a new pattern?