GaWx

The Oct 2023 PNA came in at +1.20, which is a little higher than the +0.85 to +1.15 range that I had expected. For El Niño winters since 1950, here are the Oct PNAs along with the nature of the subsequent winter’s temperatures in the SE US: 1953: +1.91 warm 2015: +1.78 warm 1979: +1.53 normal 2023: +1.20 ???? 1965: +1.14 cool 1977: +0.98 cold 1986: +0.92 cool 2014: +0.89 cool 1963: +0.69 cold 1976: +0.68 cold 1987: +0.53 cool 2009: +0.43 cold 1969: +0.35 cold 1957: +0.27 cold 1968: +0.26 cool 1951: +0.24 warm 2018: +0.21 warm 1958: +0.16 normal 1994: -0.23 mild 1997: -0.26 mild 2002: -0.65 cool 2006: -0.84 mild 1982: -0.93 normal 2004: -1.39 normal 1972: -1.50 normal 1991: -2.28 warm Like most here, I prefer BN winters. Note that 11 of 12 of the BN SE El Nino winters since 1950 occurred when the preceding Oct PNA was within +0.26 to +1.14. Based on the assumption that this clumping isn’t mainly due to randomness (I don’t think it is), I’m content with the idea that the +1.20 Oct 2023 PNA suggests an enhanced chance of a BN SE winter though I would have preferred that the Oct PNA had come in a little lower (say ~+1.00) to within the clump rather than just above the clump. The +1.20 is still good enough in my mind. Let’s say that the Oct PNA had instead come in, say, either negative, or +1.75+. I definitely wouldn’t have felt it was enhancing BN SE winter chances and probably would have felt it meant the opposite, if anything. PNA monthlies back to 1950: https://www.cpc.ncep.noaa.gov/products/precip/CWlink/pna/norm.pna.monthly.b5001.current.ascii.table

Followup to the above analyzing those QBO analogs and anti-logs: 1. El Niño DJF NAO -QBO Analogs: two had a -NAO (1986-7 and 2009-10) and the other two had a +NAO (1991-2 and 2014-5). This 50% having a -NAO is actually pretty impressive considering that -NAO winters have been pretty rare (only 6 of the last 44 (14%) winters since 1979-80). Moreover, the only two -NAO El Niño winters of the 14 (14%) El Niños since 1979-80 are the QBO analog El Niños of 1986-7 and 2009-10. The other four -NAO winters (1984-5, 1995-6, 2010-1, and 2020-1) weren’t El Niño. -QBO Anti-logs: all five had a +NAO. 2. El Niño DJF AO: -QBO Analogs: two had a -AO (1986-7 and 2009-10) and the other two had a +AO (1991-2 and 2014-5). So, 50% had a -AO vs 36% of all 14 Nino winters since 1979-80 having a -AO. So, that’s fairly close. -QBO Anti-logs: none had a -AO as 4 of the 5 (80%) were neutral and one (1994-5) had a +AO. I consider neutral AO to be between +0.5 and -0.5. Conclusions for this winter’s NAO/AO based on above regarding current QBO in combo with El Niño: - Nothing overly conclusive since sample sizes are small and the analogs were 50-50 on -NAO/-AO. - But -NAO chances this winter considering the current era of difficult to achieve -NAO winters may be enhanced somewhat due to current QBO vs where they’d be with different QBO timing. - -AO chances this winter may be enhanced some due to current QBO. - Both -NAO and -AO chances during El Niño winters would appear to be higher during analog QBO than during anti-log QBO. - I’ve in the past not been one to use the QBO much, if any, in winter predictions due to what I’ve perceived to be too much variability of analogs and the difficulty of having large enough sample sizes with similar QBO timing. After all, this QBO table goes back only to 1979. But, I will say based on this latest analysis that I feel somewhat (but not dramatically) better about the chances of a -NAO and -AO this winter than I felt before I did this. To clarify, that doesn’t mean I’m saying either is likely, especially -NAO, though I will say that -AO winters have been less difficult to attain than -NAO winters since 1979-80. I’ve been really down on the chances of a -NAO this winter (even though I’ve been leaning -NAO for Feb, alone) based on how rare they’ve been since 1979-80 along with this winter’s expected high (>100) sunspot activity. All 6 -NAO winters since 1979-80 had low sunspot activity (all had 33 or less). Maybe the QBO/El Nino combo will trump the sunspots. (Plus I know some others here feel this winter’s still ascending part of the sunspot cycle will actually be helpful for -NAO chances even with it approaching max mainly based on an external study.)

1. El Niño DJF NAO -QBO Analogs: two had a -NAO (1986-7 and 2009-10) and the other two had a +NAO (1991-2 and 2014-5). This 50% having a -NAO is actually pretty impressive considering that -NAO winters have been pretty rare (only 6 of the last 44 (14%) winters since 1979-80). Moreover, the only two -NAO El Niño winters of the 14 (14%) El Niños since 1979-80 are the QBO analog El Niños of 1986-7 and 2009-10. The other four -NAO winters (1984-5, 1995-6, 2010-1, and 2020-1) weren’t El Niño. -QBO Anti-logs: all five had a +NAO. 2. El Niño DJF AO: -QBO Analogs: two had a -AO (1986-7 and 2009-10) and the other two had a +AO (1991-2 and 2014-5). So, 50% had a -AO vs 36% of all 14 Nino winters since 1979-80 having a -AO. -QBO Anti-logs: none had a -AO as 4 of the 5 (80%) were neutral and one (1994-5) had a +AO. I consider neutral AO to be between +0.5 and -0.5. Conclusions for this winter’s NAO/AO based on above regarding current QBO combined with El Niño: - Nothing overly conclusive since sample sizes are small and the analogs were 50-50 on -NAO/-AO. - But -NAO chances this winter considering the current era of difficult to achieve -NAO winters may be enhanced somewhat due to current QBO vs where they’d be with different QBO timing. - -AO chances this winter may be enhanced some due to current QBO. - Both -NAO and -AO chances during El Niño winters would appear to be higher during analog QBO than during anti-log QBO. - I’ve in the past not been one to use the QBO much, if any, in winter predictions due to what I’ve perceived to be too much variability of analogs and the difficulty of having large enough sample sizes with similar QBO timing. After all, this QBO table goes back only to 1979. But, I will say based on this latest analysis that I feel somewhat (but not dramatically) better about the chances of a -NAO and -AO this winter than I felt before I did this. To clarify, that doesn’t mean I’m saying either is likely though I will say that -AO winters have been less difficult to attain than -NAO winters since 1979-80. But I’ve been really down on the chances of a -NAO this winter based on how rare they’ve been since 1979-80 along with this winter’s expected high (>100) sunspot activity. All 6 -NAO winters since 1979-80 had low sunspot activity (all had 33 or less). Maybe the QBO/El Nino combo will trump the sunspots.

The 30 mb QBO (which I know @nrgjeffamong others follows closely) for Oct came in about as expected at -16.98. The closest El Niño summers/autumns to 2023 (analogs) since 1979 in terms of timing of peaks and valleys are 1986, 1991, 2009, and 2014. These had their first -QBO month within two months of the preceding July and had a -QBO throughout winter. 2023’s first -QBO month was July. The furthest El Niños from 2023 (anti-logs) in terms of timing since 1979 are 1982, 1987, 1994, 2015, and 2018. These all had their first -QBO month about a year earlier than the preceding July and had a +QBO throughout winter. https://www.cpc.ncep.noaa.gov/data/indices/qbo.u30.index

The 30 mb QBO for Oct came in about as expected at -16.98. The closest El Niño summers/autumns to 2023 (analogs) since 1979 in terms of timing of peaks and valleys are 1986, 1991, 2009, and 2014. These had their first -QBO month within two months of the preceding July and had a -QBO throughout winter. 2023’s first -QBO month was July. The furthest El Niños from 2023 (anti-logs) in terms of timing since 1979 are 1982, 1987, 1994, 2015, and 2018. These all had their first -QBO month about a year earlier than the preceding July and had a +QBO throughout winter. https://www.cpc.ncep.noaa.gov/data/indices/qbo.u30.index

Although today’s EPS actually shows fewer members with a (impending) major SSW (5% vs 7%), the mean wind dropped back to 27 m/s vs 30 m/s yesterday. Lowest of any run so far is the 21 m/s of 3 days ago. Normal is 35

Probably not changing from EP to CP but changing to EP/CP cross is actually being forecasted.

The unrounded ASO ONI was +1.54, officially making this a strong Nino. Also, for the 3rd trimonth in a row, the RONI was ~0.5 cooler (near/at a record amount cooler). So, that’s a full ONI category weaker! So, odds are imho that the current low end RONI based moderate will not reach strong.

Indeed, the +AAM correlates partially to +ONI/-SOI/Nino. Thus, there’s a partial correlation for AAM at H5 in the E US to a canonical El Niño pattern. So, the current -AAM is more Ninalike than Ninolike. Does anyone have a link to daily AAM?

So, 2023’s Sep MEI was weaker than all 14 Nino’s since 1979 except 1979, 2004, 2009, 2014, and 2018. And with 1979, 2004, 2014, and 2018 all being weak, only 2009 and 2023 stand out as going strongly against the grain of the ONI/MEI correlation. (2009 does finally get above +1 Nov+.) And wasn’t 2009 arguably Modoki? If so, that makes Webb’s comment about an MEI/Modoki correlation even more questionable.

Yeah, Webb implied that and I was just repeating what he said. Looking at the graphs, MEI since 1979 was clearly highest during the three strongest Nino’s based on ONI and thus 2023 is clearly lagging those three. Do you have a link to the monthly tabular MEI values?

Yeah, I‘m just saying that Webb is saying that the atmospheric response has been “pretty strong” despite the weak MEI. As I’ve said, how one classifies the atmospheric response can vary with how one defines what that is. I’ve been saying that the SOI is one way to measure atmospheric response and it averaging -10.4 in ASO is a form of a notable atmospheric response as that is well up into -SOI range for strong El Niños since 1950. But I realize that there are other forms of A.R. like H5 outside of the tropics.

I mistakenly referred to Webb referring to a stubborn -PNA whereas it was actually TheBetterDoge saying that in a reply to Webb.

On rereading the @snowman19post, I now realize that I mistakenly attributed the “stubborn -PNA” reference to Webb when it was actually “TheBetterDoge” who referred to that in a reply to Webb. I’ll correct my earlier post to refer to that person instead. Webb is saying the atmospheric response is actually pretty strong despite the weak MEI because the MEI tends to be stronger only during Modokis and that thus there will still be good warming upcoming even with the MEI still weak.

Why is “TheBetterDoge” talking about a “stubborn” -PNA? Just because 11 of the last 14 days and the last 3 days in a row have had a -PNA? Big deal. A much bigger deal is that for only the second El Niño and 4th year since 1950, 2023 had a +PNA (+0.25+) in all of June-Oct with only 2005, 2009, and 2021 being the others. Furthermore, per GEFS there are no more -PNA days in sight with neutral to positive the next 14++ days. Included in this is a pretty strong+PNA returning midmonth. So, Nov may turn out to be the 6th +PNA month in a row. The only year since 1950 with a +PNA in all of June-Nov is 2021: https://www.cpc.ncep.noaa.gov/products/precip/CWlink/pna/norm.pna.monthly.b5001.current.ascii.table Edited for correction as TheBetterDoge rather than Webb referred to a “stubborn -PNA”.

Yeah, even the ensemble means aren’t anywhere close to reliable that far out but then again the EPS has done a far better job predicting the SPV than what’s underneath. That’s why it has my interest. Also, there actually haven’t been any wobbles as big as today’s since I started looking at it closely. Is it just a wobble? Stay tuned!

Absolutely it will wobble. But it had a large wobble today, which doesn’t have the SPV fall back below normal til ~12/10. Two days ago it did that already on 11/30. I like to follow the trends/wobbles. Sometimes the wobbles are actually pointing to a new trend. But maybe it will do a hard wobble back the other way tomorrow for all we know.

Today’s EPS 10mb 60N mean wind forecast backed off a good bit more from the weak SPV having been forecasted in recent days. Two to three days ago it had it dip (vs the 35 m/s normal) to 21 m/s. Yesterday that rose to 25 m/s. Today it is up to 30 m/s along with only ~7% of members suggesting a major SSW in early to mid Dec vs 10% yesterday and 18% 2-3 days ago. Also, the Nov and early Dec. SPV is significantly stronger than yesterday:

1. Despite the WCS PDO for Oct coming in at up at ~-1.0 from Sep’s ~-2.0 and thus my expectation that NOAA’s Oct would come in within or near -1.5 to -1.8, it actually came in still down at -2.36. That’s still a decent rise from the -2.97 of Sep but is disappointing: https://www.ncei.noaa.gov/pub/data/cmb/ersst/v5/index/ersst.v5.pdo.dat 2. The WCS daily PDO rose slightly for the 2nd straight day and was at -0.99. So, NOAA daily is probably still down near or below -2. So, there’s still a long ways to go to get anywhere close to neutral. Keep in mind that last winter had a -1.70 NOAA PDO and Oct of 2022 was -1.80 vs the -2.36 of 10/2023. So, getting a NOAA PDO for DJF to average >-1.0 will not be easy!

My point is that the seasonal models, including CANSIPS, can very easily end up way off (too warm, cold, dry, wet, etc) for winter even the day before winter starts and we’re still a month out. Regarding the PDO, it was already way down at -2.41 in Nov of 22 and it actually rose during the winter (winter averaged -1.70).

“FWIW” is the key. I’ve found these long range/climate model predictions to be of little value. I’ve seen this and other ones totally blow their winter forecasts even on Nov 30th! Here’s a great example from the 11/30/22 CANSIPS for last DJF that shows the MidAtlantic states near normal (+0.25 to +0.5C or ~+0.5 to +1F) vs 1981-2010 normals. The 1981-2010 NYC normal DJF is 35.5F. This CANSIPS was forecasting NYC to be ~36.2F. The actual winter averaged way up at 41F! So, this model’s Nov 30th run was a whopping 4.8F too cold for last winter there and similarly much too cold all around there/too cold throughout the E US. Also, it was much too warm in the W US. So, it totally blew most of the country and this was released the day before winter started: CANSIPS 11/30/22 forecast for last DJF (1981-2010 normals) Actual (1981-2010 normals): much warmer E US and much colder W US vs Nov 30th CANSIPS forecast

The WCS PDO for Nov 1st of -1.02 was actually slightly warmer than that for Oct 31st. This means that the corrective fall from the huge earlier rise may be over.

Today’s EPS backed off some (though nowhere near completely) vs yesterday’s run, which had shown the strongest indications for a weakened SPV in Dec yet. Yesterday’s had the 60N 10 mb wind dip way down to 21 m/s vs the normal of 35. Today’s has the EPS mean dip to 25 m/s, which is still significant. The prior two days of runs implied ~18% of members with actual or impending major SSWs in early to mid Dec. Today’s backed off to ~10%, which is similar to the run from three days ago. At day 45, yesterday’s had ~85% of members with a SPV weaker than climo. Today’s is ~75%.

Crestview, a perennial cold spot in N FL, had a low of 27, which ties with 1993 the coldest low for the date on record and misses just by one day for coldest so early in the season as 1993 also had a 27 on Nov 1st.

That’s why I take modeled seasonal forecasts with a grain of salt. However, the Euro Weeklies have (based on many years of data) been forecasting the stratosphere through week 6 much more accurately than the troposphere as I posted about earlier. Thus, I tend to pay extra attention to strat forecasts of the extended Euro Weeklies. The persistent forecast for a weak SPV on the Euro in Dec is quite notable as a whopping 85% of the members have a weaker than normal SPV in mid Dec.