CFS Long Range (as in Winter)

[usedtobe]

I'm not sure how weak the correlation is...even if we look at all ninos regardless of strength or weather they followed a Nina the evidence is there in the means...

The correlation is pretty weak as I noted in my talk. It's a little higher for the NAO than the AO. The highest correlation is with a PNA pattern.

[Deck Pic]

The correlation is pretty weak as I noted in my talk. It's a little higher for the NAO than the AO. The highest correlation is with a PNA pattern.

Isn't that for all three months and based on numerical data? I realize monster years skew those composites. I think my suggestion is that there is a correlation between niño and a period of blocking. Especially the later in the winter. I'd guess the correlation from say Jan 15 through Feb 28th is much higher. I think the bad 1st halfs skew the data a bit. I don't think there are more than a few ninos if that that didn't have a blocking pattern at some point, usually extended in the averages.

I do remember you mentioning the weak correlations though.

[usedtobe]

Isn't that for all three months and based on numerical data? I realize monster years skew those composites. I think my suggestion is that there is a correlation between niño and a period of blocking. Especially the later in the winter. I'd guess the correlation from say Jan 15 through Feb 28th is much higher. I think the bad 1st halfs skew the data a bit. I don't think there are more than a few ninos if that that didn't have a blocking pattern at some point, usually extended in the averages.

I do remember you mentioning the weak correlations though.

Probably so but 1972-1973 and 1991-1992 were pretty much positive through the season especially the latter in terms of AO. Also, the correlation can be low if you have a two week period of blocking in a nino season with all the rest of the season having the ao Positive.

I think that happened during the bulk of the crummy everyone along I95 corridor years based on the composite below. The snowy years in El Ninos were much different than the non-snowy ones.

I've never tried to look at ninos following ninas as I'm not sure there are enough years to dray any conclusions.

[Deck Pic]

Thanks Wes....I have theories on why all of those sucked but I may just be trying to force some weak reasoning into my "theory" to make it work..some of the examples still had good periods

51-52: too weak following Nina,, but had blocking in February

58-59: had the blocky, cold January but probably too weak to get a good storm track

72-73: too strong

91-92: too strong, especially following 2 year neutral-warm period

94-95: too strong following 4 year neutral-warm period though we had the cold, decent stretch end of January/February

97-98: too strong

06-07: too weak following Nina, but cold, -NAO FEB

the weak examples all had blocking periods...the strong (or the ones I think were too strong like 94-95) are the problem

[Deck Pic]

Thanks Wes....I have theories on why all of those sucked but I may just be trying to force some weak reasoning into my "theory" to make it work..some of the examples still had good periods

51-52: too weak following Nina,, but had blocking in February

58-59: had the blocky, cold January but probably too weak to get a good storm track

72-73: too strong

91-92: too strong, especially following 2 year neutral-warm period

94-95: too strong following 4 year neutral-warm period though we had the cold, decent stretch end of January/February

97-98: too strong

06-07: too weak following Nina, but cold, -NAO FEB

the weak examples all had blocking periods...the strong (or the ones I think were too strong like 94-95) are the problem

I just looked at the 12 ninos following nina's/negative neutral periods and regardless of strength, they were split down the middle on Dec and January as far as pos/neg nao and 10 of 12 were negative in February with one of the ones that was positive being Feb 2003

[usedtobe]

I just looked at the 12 ninos following nina's/negative neutral periods and regardless of strength, they were split down the middle on Dec and January as far as pos/neg nao and 10 of 12 were negative in February with one of the ones that was positive being Feb 2003

If the Nino rises to 1.2 and the QBO goes westerly by winter, then with the solar flux increasing we have a decent chance at a decent back loaded winter. I'm not all that comfortable about what ninos following ninas mean since the ninas themselves can be different and we have a limited number of years. Still I'd 10 out of 12 giving us a negative NAO in Feb than not. MY own view is that a weakly negative AO might not mean much.

[Bob Chill]

I just looked at the 12 ninos following nina's/negative neutral periods and regardless of strength, they were split down the middle on Dec and January as far as pos/neg nao and 10 of 12 were negative in February with one of the ones that was positive being Feb 2003

The thing about Feb 2003 is that + reading for the month was mostly from the strong + during the first week of the month. Jan had a sick stretch of -nao with reading down below -3 and then it rebounded at the end of the month that lasted through mid Feb. The last half of Feb 2003 was in neg territory and that lasted all the way through the first 8 days of March.

I guess what I'm trying to say is that from a practical standpoint, Feb 03 had a very favorable NAO so the + reading for the month doesn't tell the whole story. The entire stretch of Jan - 1st week of March in 03 was really really good irt the NAO. The persistence was definitely biased towards negative with a few short lived strong + periods.

I'd considering sacrificing one of my yard bunnies for another Jan-Mar 03. They eat all my damn vegetable plants anyway so I might sacrifice for other reasons.

[Deck Pic]

If the Nino rises to 1.2 and the QBO goes westerly by winter, then with the solar flux increasing we have a decent chance at a decent back loaded winter. I'm not all that comfortable about what ninos following ninas mean since the ninas themselves can be different and we have a limited number of years. Still I'd 10 out of 12 giving us a negative NAO in Feb than not. MY own view is that a weakly negative AO might not mean much.

Yeah. Our sample size is pretty bad. So there is no way I can really glean any sort of rule. Just maybe a tendency. Of course then we need the timing and the setup and all the details to line up.

Re: the QBO. Wont it still be easterly or neutral by winter?

[donsutherland1]

In my weenie mind I always assumed a mega-drought in the summer and fall led to a wet winter. Rubber-band and all that.

2/3 of the 30 driest summers wound up also being among the 50 driest winters (CONUS). Whether or not fall was dry was not a big factor.

[donsutherland1]

Some numbers on the ENSO and AO.

December, January, February (1950-2012) when the ENSO R3.4 Anomaly was +0.50°C or above:

Mean AO: -0.489

- 57% days had an AO < 0

- 38% days had an AO of -1 or below

- 13% days had an AO of -3 or below

- 43% days had an AO > 0

- 25% days had an AO of +1 or above

- 4% days had an AO of +3 or above

Ratios:

AO- : AO+: 1.3 days

AO -1 or below : AO +1 or above: 1.5 days

AO -3 or below : AO +3 or above: 3.2 days

December, January, February (1950-2012) when the ENSO R3.4 Anomaly was -0.50°C or below:

Mean AO: -0.235

- 54% days had an AO < 0

- 34% days had an AO of -1 or below

- 8% days had an AO of -3 or below

- 46% days had an AO > 0

- 26% days had an AO of +1 or above

- 4% days had an AO of +3 or above

Ratios:

AO- : AO+: 1.2 days

AO -1 or below : AO +1 or above: 1.3 days

AO -3 or below : AO +3 or above: 1.9 days

In sum, the incidence of AO- and AO+ days was similar in El Niño or La Niña winters. Severe blocking (AO -3 or below) was more than three times more likely during El Niño winters than La Niña winters.

[Deck Pic]

Some numbers on the ENSO and AO.

December, January, February (1950-2012) when the ENSO R3.4 Anomaly was +0.50°C or above:

Mean AO: -0.489

- 57% days had an AO < 0

- 38% days had an AO of -1 or below

- 13% days had an AO of -3 or below

- 43% days had an AO > 0

- 25% days had an AO of +1 or above

- 4% days had an AO of +3 or above

Ratios:

AO- : AO+: 1.3 days

AO -1 or below : AO +1 or above: 1.5 days

AO -3 or below : AO +3 or above: 3.2 days

December, January, February (1950-2012) when the ENSO R3.4 Anomaly was -0.50°C or below:

Mean AO: -0.235

- 54% days had an AO < 0

- 34% days had an AO of -1 or below

- 8% days had an AO of -3 or below

- 46% days had an AO > 0

- 26% days had an AO of +1 or above

- 4% days had an AO of +3 or above

Ratios:

AO- : AO+: 1.2 days

AO -1 or below : AO +1 or above: 1.3 days

AO -3 or below : AO +3 or above: 1.9 days

In sum, the incidence of AO- and AO+ days was similar in El Niño or La Niña winters. Severe blocking (AO -3 or below) was more than three times more likely during El Niño winters than La Niña winters.

Thanks Don

[usedtobe]

Yeah. Our sample size is pretty bad. So there is no way I can really glean any sort of rule. Just maybe a tendency. Of course then we need the timing and the setup and all the details to line up.

Re: the QBO. Wont it still be easterly or neutral by winter?

July will mark the 12th straight easterly qbo month. the last two lasted 14 and 15 months but they can last longer and the one before it last 18 months. We probably switch to westerly sometime around OCt or November. With the solar cycle being on the rise, a westerly qbo might be better at getting a late season AO and NAO through a stratospheric warming event than an easterly one. That said, the AO and NAO are sort of a crap shoot this far in advance.

[aldie 22]

After being on here for so many years I find myself getting less excited each year with the lead up to winter...you just resign yourself that our area generally has a better chance at lackluster vs. blockbuster...no such thing as normal as far as I have seen. I'm hopeful that this won't be a repeat of last year but sometimes you just never know....I still love me some snow and cold though

[WxUSAF]

1983,1993, 2003 = 2013 big snow. 1 time it's a fluke, 2 your on to something & 3 is a pattern.

[North Balti Zen]

When you quote him...it defeats the purpose of the "ignore" function for some of us...

[ORH_wxman]

New Euro SIPS (still not out publically on the ECMWF.int site but should be soon) has a mean around 1.2C or so for the Nino on the monthly peak...which for trimonthly would probably be about 2 tenths lower...so its definitely leaning toward a weak Nino more than it has been from the June and May runs. That is not as favorable for DC as a 1.3-1.8 type peak.

However, the Euro SIPS has two clusterings of values...the larger cluster is around 0.9C and the smaller cluster is around 2.1C...very far apart in their values. This shows a lot of uncertainty in the development of this Nino. I am definitely leaning toward the larger cluster at the moment, but its very possible this could strengthen rapidly in latter August and September once this unfavorable pattern wanes. We'll just have to wait and see. I think we'll have a pretty good idea of whether we can eliminate a peak >1.5C or not by the end of August.

[WxMidwest]

One would think it would be hard to get a Nino to peak out above a 2.0C SST anomaly, during a -PDO cycle... Concensus Weak, maybe low-end moderate at most. Wouldn't be surprised if it tended to be more of a Modoki/Central Pacific Nino instead of your traditional...

[Deck Pic]

New Euro SIPS (still not out publically on the ECMWF.int site but should be soon) has a mean around 1.2C or so for the Nino on the monthly peak...which for trimonthly would probably be about 2 tenths lower...so its definitely leaning toward a weak Nino more than it has been from the June and May runs. That is not as favorable for DC as a 1.3-1.8 type peak.

However, the Euro SIPS has two clusterings of values...the larger cluster is around 0.9C and the smaller cluster is around 2.1C...very far apart in their values. This shows a lot of uncertainty in the development of this Nino. I am definitely leaning toward the larger cluster at the moment, but its very possible this could strengthen rapidly in latter August and September once this unfavorable pattern wanes. We'll just have to wait and see. I think we'll have a pretty good idea of whether we can eliminate a peak >1.5C or not by the end of August.

[mitchnick]

gotta' like what the CFS has been showing now for a month+ and how the CFS2 is evolving toward a colder look as well

time to start buying nat gas futures

[donsutherland1]

gotta' like what the CFS has been showing now for a month+ and how the CFS2 is evolving toward a colder look as well

time to start buying nat gas futures

The trend on the CFSv2 is encouraging. Should the winter be blocky, the eastern U.S. will have a good chance of averaging cooler than normal. The area covered by cold anomalies will probably be less expansive than that for winters 2009-10 or 2010-11, as the blocking for those two winters was extreme. Perhaps cold anomalies will cover a similar area +/- a modest amount as one saw during the 2002-03 winter. Whether the drought expands eastward could have an impact on precipitation and opportunities for snowfall. Of course, a lack of blocking would result in a much warmer outcome, but for now the model trends are encouraging.

[Ellinwood]

gotta' like what the CFS has been showing now for a month+ and how the CFS2 is evolving toward a colder look as well

time to start buying nat gas futures

Let's take a look at CFS verifications from say, oh February for the late spring into early summer:

Now let's compare it to the actuals...

May-Jun-Jul (to date):

Much, much too cold across most of the CONUS and central Canada, and I hardly think that Jun-Jul-Aug is going to come in near normal across most of North America.

The CFS has been consistently too cold for awhile, especially once you get more than a couple of months into the future. I really wouldn't be giving the CFS any weight WRT the actual forecast anomalies.

Now, the CFSv2 I'd give some weight to, but I definitely wouldn't concentrate on it.

[Deck Pic]

Kind of an interesting observation.

Re: The dynamical model averages(statistical models arent very good) in the last 4 Niño events.

2002, 2004, 2006, 2009 all had exactly a .2 degree cold bias for DJF tri monthly in the July Averages. Probably coincidence but would make

This niño a 1.1 which is still risky territory but could be decent.

NASA and COLA CCSM3 (which is a similar coupled model to the CFS) are warm biased models so I wouldnt put too much stock in their exact output right now. When you account for their bias, low 1's seems reasonable. I don't think there is a lot of current evidence that niño will go above 1.2 but I hope I am wrong. That is still risky territory for good snow though at least one month or 2 should be cold.

[donsutherland1]

Now, the CFSv2 I'd give some weight to, but I definitely wouldn't concentrate on it.

I agree. If I recall correctly, the CFS will be discontinued after October. Overall, I don't believe that model has done well for the most part. The CFSv2 has fared better.

[mitchnick]

Let's take a look at CFS verifications from say, oh February for the late spring into early summer:

Now let's compare it to the actuals...

May-Jun-Jul (to date):

Much, much too cold across most of the CONUS and central Canada, and I hardly think that Jun-Jul-Aug is going to come in near normal across most of North America.

The CFS has been consistently too cold for awhile, especially once you get more than a couple of months into the future. I really wouldn't be giving the CFS any weight WRT the actual forecast anomalies.

Now, the CFSv2 I'd give some weight to, but I definitely wouldn't concentrate on it.

otoh, CFS1 hit with darn good accuracy for a seasonal forecast last year in this time frame as well as 09/10 winter; here are its predictions for winters of 09/10 and 11/12 from the first week of August

[Ellinwood]

otoh, CFS1 hit with darn good accuracy for a seasonal forecast last year in this time frame as well as 09/10 winter; here are its predictions for winters of 09/10 and 11/12 from the first week of August

"darn good accuracy" may be a subjective term... I prefer "okay":

Don't get me wrong, it's okay at picking out the pattern itself... I'm just saying I have zero confidence in the actual anomalies it shows.

EDIT: Granted, it did a real sh**ty job at picking out the pattern for this summer, so it's kind of hit or miss WRT the pattern as well.

[Ellinwood]

Also, for the sake of not cherry-picking verifications, here's the winter 2010-11 forecast from August and the verification to go with the other winters:

Not so good.

[griteater]

Kind of an interesting observation.

Re: The dynamical model averages(statistical models arent very good) in the last 4 Niño events.

2002, 2004, 2006, 2009 all had exactly a .2 degree cold bias for DJF tri monthly in the July Averages. Probably coincidence but would make

This niño a 1.1 which is still risky territory but could be decent.

NASA and COLA CCSM3 (which is a similar coupled model to the CFS) are warm biased models so I wouldnt put too much stock in their exact output right now. When you account for their bias, low 1's seems reasonable. I don't think there is a lot of current evidence that niño will go above 1.2 but I hope I am wrong. That is still risky territory for good snow though at least one month or 2 should be cold.

Good analysis. A blend of the EUROSIP and the Dynamical Model Averages is probably the best we've got with respect to ENSO projections.

[Bob Chill]

Even a nino in the 1 - 1.3 range isn't something that's concerning. At least I don't think it is. At the very least it opens the door for potential and we really need that coming off the heels of a multi-year cold enso.

Probably going to get going later rather than soon but at some point we will see an active stj. Then it comes down to the setup. I see no evidence that points towards another crappy +AO/NAO type of winter. We'll have our chances this winter. It's foolish to ever predict a big winter in our area because the odds are so bad but there is no reason to believe we can't go slightly above climo with at least one 6"+ storm.

[mitchnick]

"darn good accuracy" may be a subjective term... I prefer "okay":

Don't get me wrong, it's okay at picking out the pattern itself... I'm just saying I have zero confidence in the actual anomalies it shows.

EDIT: Granted, it did a real sh**ty job at picking out the pattern for this summer, so it's kind of hit or miss WRT the pattern as well.

although I didn't say it, I was being Mid-Atlantic centric wrt the CFS seasonal forecasts of 09/10 and 11/12

it still did OK with the general pattern in those years, though it did blow 10/11

I expect we could all agree on the general accuracy of the seasonal models at this time range and the CFS has had its share of big misses with the rest of them, but its forecast at this range for 2 out of the last 3 winters for our region have been pretty good imho, so I'm liking what I see, which is what I said in my post on Page 4

moreover, considering we're talking a NINO this year, in one strength or another, and the trend on the CFS2, I think we do have better than equal chances of finally waking from last winter's nightmare

[Ellinwood]

although I didn't say it, I was being Mid-Atlantic centric wrt the CFS seasonal forecasts of 09/10 and 11/12

it still did OK with the general pattern in those years, though it did blow 10/11

I expect we could all agree on the general accuracy of the seasonal models at this time range and the CFS has had its share of big misses with the rest of them, but its forecast at this range for 2 out of the last 3 winters for our region have been pretty good imho, so I'm liking what I see, which is what I said in my post on Page 4

moreover, considering we're talking a NINO this year, in one strength or another, and the trend on the CFS2, I think we do have better than equal chances of finally waking from last winter's nightmare

lol, that's a small difference in perspective between us

As far as the CFS is concerned, there really isn't much (if any) value in using that model anymore. It's going down in October, anyway... might as well just stick with the CFSv2 and other models that aren't the CFS (at least those that aren't as bad or worse as the CFS). IMO, using the CFS 4-6+ months in advance is like using the JMA for a winter storm 7 days out (or at any time, for that matter ).