2024-2025 La Nina

[donsutherland1]

1 hour ago, Terpeast said:

The fact that the CFS caught it early lends more credibility to that model.

I agree. It was even more surprising, as the CFSv2 has been struggling with the ENSO.

[snowman19]

2 hours ago, GaWx said:

 Thanks, Don. Not what I want to see. Thus I’m enjoying the current cold pattern before it ends. 
I’ll be watching to see if NYC can get >3” of snow this month per @bluewaveindicator for rest of season’s snow prospects there. Do you have a prediction of whether or not NYC’s snow will likely exceed 3” this month?

The day before Thanksgiving someone had tweeted the actual data on La Nina’s and cold-neutrals in the NYC area, going back to 1945. Long and short of it, they found that the overwhelming majority of the time, when December produced above normal snowfall, the rest of the season (Jan-Feb-Mar) went on to see average above normal snowfall. The exact opposite was true for Decembers that produced below average snowfall. So it definitely backs up @bluewave findings 

[mitchnick]

51 minutes ago, GaWx said:

I’ve been emphasizing how the Euro had 12/2-8 as much too warm not long ago. So, these outlooks are very far from being reliable. But at the same time I’d still rather them not trend warmer like most later weeks did yesterday.

 Aside: NG now -5%/low of day.

We'll get the Euro monthly tomorrow if I'm not mistaken. I think it comes out on the 4th of the month these days??? So we'll have 1 more piece of guidance to fight over! Lol

[GaWx]

25 minutes ago, mitchnick said:

We'll get the Euro monthly tomorrow if I'm not mistaken. I think it comes out on the 4th of the month these days??? So we'll have 1 more piece of guidance to fight over! Lol

The NG market is now down a whopping 15% vs a week ago. I learned quite awhile back that this market is smart (as @Stormchaserchuck1knows well) and is thus a great barometer of expectations of temperatures in the E US in the decently predictable medium range. Being down 15% is saying a lot as regards how things looked in the medium range one week ago, when near term cold looked like it might last through the first half of Dec.

[mitchnick]

19 minutes ago, GaWx said:

The NG market is now down a whopping 15% vs a week ago. I learned quite awhile back that this market is smart (as @Stormchaserchuck1knows well) and is thus a great barometer of expectations of temperatures in the E US in the decently predictable medium range. Being down 15% is saying a lot as regards how things looked in the medium range one week ago, when near term cold looked like it might last through the first half of Dec.

Legalized gambling really. OK, I  guess, as long as you know that going into it, but it's not for me. You're really trying to predict wx models and not the weather, because by the time the weather becomes certain, you've already lost or made your money.

[roardog]

3 minutes ago, mitchnick said:

Legalized gambling really. OK, I  guess, as long as you know that going into it, but it's not for me. You're really trying to predict wx models and not the weather, because by the time the weather becomes certain, you've already lost or made your money.

The reality is that most of the people posting in this thread probably know what type of pattern the models are showing in the medium or long range before the natural gas market does. I don’t really think it’s telling us anything we don’t already know. 

[GaWx]

On 11/27/2024 at 10:33 PM, so_whats_happening said:

I expect the PDO to be under -3 for the month of November which is a good thing to see.

 

On 11/27/2024 at 11:53 PM, GaWx said:

 It looks to me like it will be close. Right now I have Nov MTD near the -2.9 to -3.3 range. My best guess is for -3.0 to -3.2. Jul and Aug were in the -2.9s. It appears that Nov was slightly more negative than those two, but the correlation of NOAA to WCS from month to month has been far from perfect. Thus, another in the -2.9s seems like a legit possibility even though I favor sub -3.0 right now. We’ll know pretty soon. If Nov is sub -3.0, it would be only the second time since 1851 for three sub -3 in a row. The only other time was in 1894.

 Nov NOAA PDO just came in about where I expected (based on adjusting WCS’ PDO) at -3.24 (rise of 0.54 from Oct). This compares to the WCS’ Nov of ~-2.10. So, this is only the 2nd time since 1851 with three sub -3 PDO months in a row. The other time was in 1894 (6 months earlier that year than 2024). We actually only barely missed 7 months in a row of sub -3 as the last 7 have all been sub -2.90:
 

-2.99 -3.16 -2.99 -2.91 -3.54 -3.78 -3.24

https://www.ncei.noaa.gov/pub/data/cmb/ersst/v5/index/ersst.v5.pdo.dat

 Current daily NOAA PDOs are likely in the low -3 range.

@snowman19

[GaWx]

34 minutes ago, roardog said:

The reality is that most of the people posting in this thread probably know what type of pattern the models are showing in the medium or long range before the natural gas market does. I don’t really think it’s telling us anything we don’t already know. 

NG largely ignores the long range because it is so highly unreliable. That’s why it focuses on the medium range. It reacts very soon after when medium range runs, especially EPS and GEFS, are released.

[40/70 Benchmark]

5 hours ago, PhiEaglesfan712 said:

Didn't we get one in 2015?

Yea...1972, 1982, 1997, 2015......several more years and we're due.

[40/70 Benchmark]

3 hours ago, donsutherland1 said:

The guidance has turned milder. The CFSv2 was hinting at this for several days and the other guidance fell in line. Often the guidance rushes such transitions, but not this time. I suspect much of the rest of December could be mild in the eastern half of the CONUS (still sufficiently cold at times for snow in the Great Lakes Region) but reduced snowfall chances for NYC and southward. Even the current colder period is on track to be milder than what had previously been shown on the guidance.

I was a bit nervous for a spell, but feel better about my December call now.

[leo2000]

6 minutes ago, 40/70 Benchmark said:

Yea...1972, 1982, 1997, 2015......several more years and we're due.

Could be anytime as 1972, 1982 was only 10 years apart. Wasn't 2013-2014 winter a good positive PDO too?. 

[snowman19]

Looks like yesterday’s GEFS may have been onto something

[40/70 Benchmark]

8 minutes ago, leo2000 said:

Could be anytime as 1972, 1982 was only 10 years apart. Wasn't 2013-2014 winter a good positive PDO too?. 

I don't think its imminent given the magnitude of last year's El Nino.

[PhiEaglesfan712]

10 minutes ago, leo2000 said:

Could be anytime as 1972, 1982 was only 10 years apart. Wasn't 2013-2014 winter a good positive PDO too?. 

The winter of 2013-14 itself wasn't a +PDO, but was one that trended in that general direction. 2014-15 was the +PDO winter, as was the super el nino of 2015-16.

2016-17 was pretty much the opposite of 2013-14, in regards to the PDO.

[qg_omega]

This week is a blip, thoughts most had in early November will verify

[MJO812]

14 minutes ago, qg_omega said:

This week is a blip, thoughts most had in early November will verify

It doesn't look like a long lasting warmup

[bluewave]

2 hours ago, mitchnick said:

Here's even a better example. 2 weeks ago, 11/18 forecast, for the week of 122-12/9 on top. Yesterday's forecast for the same week that's upon us so there's no question how close the prediction will be. 

Your colder example is the exception to the rule which will eventually happen from time to time. But most of the time the models adjust warmer as we get closer we get to the forecast period. The colder pattern than originally forecast around the start of December was the result of the models adjusting to the +AAM rise. But at the same time they underestimated the forcing currently in MJO 4 going over +2.00. Stronger forcing in that part of the world accelerates the Pacific Jet and weakens the -EPO. So you had one event strengthening the -EPO and another weakening it. The height falls showing up near Alaska around the 7th with a faster jet than forecast weren’t there previously when the models were underestimating the forcing near the Maritime Continent. This process lead to the quick change in the week 2 forecast to much warmer than previous runs. This is why I was mention earlier back in this thread that the EPS was probably underestimating the Pacific Jet.

[LakePaste25]

Most of Canada stays cold through the entire run. While we will get warmer, it is a far cry from last December when we had the pacific jet extended all the way to the west coast, and it was just torching our cold source region with mild pacific air. Our source region for cold will get some nice chilling hours, which means any favorable pattern shift will easily bring cold air south later in the month. 

[GaWx]

 The brand new Euro Weeklies (which start Dec 9th) are mainly warmer than yesterday’s run in the E US in week by week comparisons for the last 3 full weeks of Dec. They are all mainly only slightly AN (~+2 to +3 vs 1991-2020 mean) meaning they’re not showing a blowtorch and still may not fully negate the very cold early Dec if they were to verify well.

 The first 2 weeks of Jan in the NE US are about the same as yesterday and are NN.

[snowman19]

It doesn't look like a long lasting warmup

Ummmmm yea it sure does. If the EPS and GEFS are correct, the EPO floodgates open up and it will scour all the arctic air out of Canada. Won’t happen immediately, but it will happen. This pattern looks to have staying power. Don’t believe me? Fine. Here’s Eric Webb and Earthlight:

[Maestrobjwa]

4 minutes ago, snowman19 said:

Ummmmm yea it sure does. If the EPS and GEFS are correct, the EPO floodgates open up and it will scour all the arctic air out of Canada. Won’t happen immediately, but it will happen. This pattern looks to have staying power. Don’t believe me? Fine. Here’s Eric Webb and Earthlight:

 

 

 

 

I'm probably way outta of pocket for saying this, and it may be a blatantly obvious answer, but...do you prefer cold weather or warm weather?

[GaWx]

37 minutes ago, bluewave said:

Your colder example is the exception to the rule which will eventually happen from time to time. But most of the time the models adjust warmer as we get closer we get to the forecast period. The colder pattern than originally forecast around the start of December was the result of the models adjusting to the +AAM rise. But at the same time they underestimated the forcing currently in MJO 4 going over +2.00. Stronger forcing in that part of the world accelerates the Pacific Jet and weakens the -EPO. So you had one event strengthening the -EPO and another weakening it. The height falls showing up near Alaska around the 7th with a faster jet than forecast weren’t there previously when the models were underestimating the forcing near the Maritime Continent. This process lead to the quick change in the week 2 forecast to much warmer than previous runs. This is why I was mention earlier back in this thread that the EPS was probably underestimating the Pacific Jet.

 In other words, the models have a cold bias in the E US. They adjust warmer more often than adjusting colder as we get closer. This has been the case for probably 8+ years. The CMC ensemble actually has had the coldest bias at 2 meters of the big 3. I used to subscribe to a company that analyzed biases in detail. That’s how I know this. The W US is a different story. The bias is much closer to neutral with a few cities sometimes actually showing a warm bias due to overdone W ridging/underdone W troughing.

[bluewave]

42 minutes ago, GaWx said:

 In other words, the models have a cold bias in the E US. They adjust warmer more often than adjusting colder as we get closer. This has been the case for probably 8+ years. The CMC ensemble actually has had the coldest bias at 2 meters of the big 3. I used to subscribe to a company that analyzed biases in detail. That’s how I know this. The W US is a different story. The bias is much closer to neutral with a few cities sometimes actually showing a warm bias due to overdone W ridging/underdone W troughing.

The ridge in the East has been one of the most under-forecast features beyond 7 days going back to December 2015 which leads to warmer temperatures than originally forecast. My guess is that this has something to do with the models underestimating long range convection near the Maritime Continent. Since forcing in that region leads to higher 500mb heights and temperatures in the East.

11-15 EPS forecast significantly underestimated Maritime Continent forcing

 

Verification much stronger forcing

 

 

 

 

[mitchnick]

I'm sorry, I  don't believe this model bias talk.  If there was a proven bias, we would be hearing about it from all the Mets as the reason for their blown forecasts. Anyway, the models are run so many times a day, what runs are you going to use and how often do you check? And is this bias on all the models all the time or some of the models some of the time? It's a moving object that never stops.

There's no doubt the models are flawed, but using 1 particular time on  run vs another 1 particular time on a run to prove a point is cherry picking where I come from. Especially if you can take a different time on the same run and show the model has gotten colder.

Frankly, I'm still trying to figure out the point of the exercise. I have my thoughts, of course.

[bluewave]

35 minutes ago, mitchnick said:

I'm sorry, I  don't believe this model bias talk.  If there was a proven bias, we would be hearing about it from all the Mets as the reason for their blown forecasts. Anyway, the models are run so many times a day, what runs are you going to use and how often do you check? And is this bias on all the models all the time or some of the models some of the time? It's a moving object that never stops.

There's no doubt the models are flawed, but using 1 particular time on  run vs another 1 particular time on a run to prove a point is cherry picking where I come from. Especially if you can take a different time on the same run and show the model has gotten colder.

Frankly, I'm still trying to figure out the point of the exercise. I have my thoughts, of course.

If you go back in these various threads including the ones in the NYC Metro forum you will see me positing model biases all the time. This is how I have been able to identity where the potential errors were to the forecasts going back years in these threads. I apply these corrections manually. If you can train the models with AI to adjust for these biases, then my guess is that the forecasts for weeks 2 and 3 would greatly improve. This could also lead to better seasonal forecasts beyond the 1 month range.

[donsutherland1]

2 hours ago, 40/70 Benchmark said:

I was a bit nervous for a spell, but feel better about my December call now.

Your call looks good. The latest ECMWF weeklies and CFSv2, which show persistent warmth for an extended period of time, raise questions about assumptions in some circles e.g., BAMWX, about this winter's seeing a predominantly cold pattern. The seasonal guidance suggested otherwise, even as month-to-month variability is likely.

[GaWx]

38 minutes ago, mitchnick said:

I'm sorry, I  don't believe this model bias talk.  If there was a proven bias, we would be hearing about it from all the Mets as the reason for their blown forecasts. Anyway, the models are run so many times a day, what runs are you going to use and how often do you check? And is this bias on all the models all the time or some of the models some of the time? It's a moving object that never stops.

There's no doubt the models are flawed, but using 1 particular time on  run vs another 1 particular time on a run to prove a point is cherry picking where I come from. Especially if you can take a different time on the same run and show the model has gotten colder.

Frankly, I'm still trying to figure out the point of the exercise. I have my thoughts, of course.

The cold E US bias is very well known because it is so obvious after analysis of numerous verifications and it has been going on for many years. I’m talking based on a compilation of all of the runs from every day for major cities throughout the US done scientifically by pro mets. It has nothing to do with cherry picking. If you still have any doubt, subscribe to Maxar. You and others would learn a lot from them. They have many pro mets on staff. Their forecasts always take these biases into account making them “bc” (bias corrected). They are not cold hypesters like JB nor cold Debby Downers. Maxar makes money off of their reputation for accuracy, not off of hype or down-playing. They forecast with emotion left out. Weenieism or emotional based bias is strongly frowned upon.

 When Chris and I talk about bias, we mean the average discrepancy between forecasted and verified actual.

[donsutherland1]

5 minutes ago, GaWx said:

The cold E US bias is very well known because it is so obvious after analysis of numerous verifications and it has been going on for many years. I’m talking based on a compilation of all of the runs from every day for major cities throughout the US done scientifically by pro mets. It has nothing to do with cherry picking. If you still have any doubt, subscribe to Maxar. You and others would learn a lot from them. They have many pro mets on staff. Their forecasts always take these biases into account making them “bc” (bias corrected). They are not cold hypesters like JB nor cold Debby Downers. Maxar makes money off of their reputation for accuracy, not off of hype.

 When Chris and I talk about bias, we mean the average discrepancy between forecasted and verified actual.

Model bias data can also be found here: https://frontierweather.dtn.com/verification.html

 

[GaWx]

11 minutes ago, donsutherland1 said:

Model bias data can also be found here: https://frontierweather.dtn.com/verification.html

 

Interesting! The DTN bias graphs for GEFS/EPS you linked me for days 6-15 averaged over the last 90 days are surprisingly showing a warm bias in the NE US. I wonder why.

 If I get a chance, I’m going to email my Maxar pro met contact. Maybe he can give me an update on the week 2 biases.

 

[bluewave]

31 minutes ago, GaWx said:

Interesting! The DTN bias graphs for GEFS/EPS you linked me for days 6-15 averaged over the last 90 days are surprisingly showing a warm bias in the NE US. I wonder why.

It was probably due to the ridge axis setting up over the Upper Midwest and Canada. So the models underestimated the temperatures under the strongest ridging. The ridge most other times has been located further east. So this has been where the typical cold bias has been. Unfortunately, DTN only posts the 90 day averages. So we don’t get to see the biases for specific patterns which produce a ridge in the East. Pattern dependent biases is the next level but they may not have the computer tech to pull off such specific data. I have noticed when the ridge sets up over the East this is when the cold biases occur. Shift it back to the upper Midwest and this is where the cold biases occur. But the northeast when averaged out over the last 10 years shows a consistent longer range bias beyond the what the shorter 90 day average shows.

Most recent 10 year 500 mb height anomaly stronger in East which is a reversal from previous 10 year period