GaWx

I was just going to respond with this link to monthly NAO back to 1950. I’ve already back tested it and have found there’s been no correlation of a -NAO in Oct to a -NAO in the subsequent DJF. Anyone could look at this data and determine this quickly. If one would look in the table, they’d see many examples of a -NAO in Oct followed by a neutral NAO or +NAO in DJF. This is similar to the lack of correlation between a summer -NAO and the subsequent winter -NAO. Of course there being no correlation of Oct -NAO and DJF -NAO doesn’t at all mean there can’t be a -NAO for DJF. It just means that having a -NAO in Oct doesn’t at all increase that chance.

That’s what I thought. Hoping the trend is wrong, which is quite possible since model accuracy this far out is limited for all models.

Where are you saying that the model trended colder? Please be more specific. My take is that you’re correct for most of the W half of the CONUS, but not for most of the E half, which actually appears to me to be warmer per this new run. Of course, this is just a model prog, which has limitations on accuracy. As one who’d prefer the coldest centered in the E US, I hope the model trend on this run turns out to be wrong, which I feel is a reasonable hope, especially considering how cold OCT may end up in the SE US (looking at past chilly Octobers).

I look at it differently. We all have different takes, preferences, etc. Here’s mine: 1. I prefer BN in all seasons. 2. I don’t believe so much in the idea of “saving” BN as in some cases it is actually the opposite: there’s a correlation and thus BN may predict subsequent BN. 3. During El Niño in the SE, there’s evidence of a partial correlation between a BN Oct and a BN DJF. Check out these 9 KSAV BN (more than 2 BN) Nino Octobers for example: - 1876, 1885, 1930, 1957, 1958, 1976, 1977, 1987, 2006 Of these nine, seven (78%) of the subsequent winters were BN at KSAV: 1876-7, 1885-6, 1930-1, 1957-8, 1976-7, 1977-8, 1987-8 The other two (22%), 1958-9 and 2006-7, were NN None (0%) were AN. These 11 El Niño Octs were AN at KSAV -1884, 1911, 1918, 1919, 1939, 1941, 1951, 1969, 1986, 2002, 2018 Subsequent winters: 3 (27%) BN: 1939-40, 1969-70, 2002-3 6 (55%) NN: 1884-5, 1911-2, 1918-9, 1919-20, 1941-2, 1986-7 2 (18%) AN: 1951-2, 2018-9 So after BN Octs El Niño winters were 78% BN, 22% NN, and 0% AN. Thus, BN El Niño winters have been favored after BN Octobers. So after AN Octs El Niño winters were 27% BN, 55% NN, and 18% AN. Thus, NN El Niño winters have been favored over each of the other categories after AN Octobers. All 7 El Nino Octobers colder than 65F were followed by BN winters.

If the 12Z GEFS suite were to verify pretty closely, the SE would have a decent shot at one of the coldest Octobers of the last few decades. Let’s say possibly 90% percentile cold in relation to the last 30 years or so. Going to be interesting to see if this model is onto something.

I don’t know what the 0% significance means on any of the maps. I got these maps from this website: https://www.daculaweather.com/4_mjo_phase_forecast.php#thumb

Bastardi said in his most recent Saturday Summary that the best shot for a cold E US winter would be for 3.4 anomalies to be higher (central) than that for Nino 1+2 (east) but also higher than 4 (west). He implied that forcing centered on Nino 4 would favor MJO phase 6, which is one of the three warmest phases in DJF overall in the E US (see maps in image below) whereas forcing centered on Nino 3.4 would favor colder E US phases (8 and 1). It looks to me like Nino 4 could also favor phase 7, which isn’t as warm in the E US as phase 6 though it still favors mildness:

For the 2nd month in a row after being mainly the other way around for awhile, ERSST came in slightly warmer than OISST in Sep (+1.59 vs +1.53). For Aug/Sep, ERSST came in 0.05/0.06 warmer than OISST.

Though it is mathematically possible based on the past for a rapid enough rise to allow for a neutral to slightly +PDO averaged out for DJF, especially with it being a solid El Niño, a more realistic possibility is only rising to the -0.50 to -1.00 area. But that would still mean a rapid PDO rise between Sep and DJF to only a modest -PDO, which would appear to be enough to correlate more with higher cool to cold E US winter chances than mild or warm E US winter chances. I don’t feel a rise to a neutral or +PDO would be what’s required.

In your outlook, you said: “I've decided to take a blend of 1952-1953, 1956-1957, 1972-1973 & 2009-2010.” Also, following that, the three maps you showed included 1956-7 instead of 1957-8 (H5, temperatures, and precip) in the Nov-Mar blend calculation. Obviously, you meant 1957-8 (strong CP El Niño), which I see you mentioned earlier several times, instead of 1956-7 (cold neutral).

Essentially, almost analogous to constructive and destructive interference in sound. Another relationship, from the University of Alaska, Fairbanks webpage The Pacific Decadal Oscillation also asserts that -One thing the severely -PDO isn’t interfering with is the solid -SOI of the last 2 months -A sub -2 Sep PDO has been followed by several cool to cold E US winters: - 1892-3: DJF PDO rose to -0.77 - 1894-5: DJF PDO rose to -0.89 - 1916-7: (cool NE): DJF PDO still down at -1.96 - 1933-4: DJF PDO rose to -0.32 (very cold Feb NE US) - 1950-1: DJF PDO rose to -1.26 (cool Mid Atlantic/SE) - 2010-1: DJF PDO still down at -1.77 - 1920-1: near normal; DJF PDO -1.11 - 1955-6: near normal; DJF PDO -2.25 - 1999-0: near normal; DJF PDO -1.80 - 2011-2: warm; DJF PDO -2.20 - 2012-3: mild; DJF PDO -1.28 - 2022-3: warm; DJF PDO -1.70

The Sep ERSST based PDO came in at a whopping -2.94. This is the 3rd lowest Sep PDO on record going way back to 1854! The only lower Sept ones were the -2.99 of 2012 and the -2.97 of 1933. Both rose rapidly in Oct and Nov and reached -0.66/-0.49 in November: https://www.ncei.noaa.gov/pub/data/cmb/ersst/v5/index/ersst.v5.pdo.dat

It is a combo of the +2 PNA (peaking on that day) and a -1 NAO coinciding.

GEFS is still suggesting high chance for a +PNA for October as a whole with a likely +2+ peak in the dailies coming early next week. Hoping Oct doesn’t come in with too high of a +PNA, however:

Actually, in addition to Sep of 2004, it is Sep of 2015 rather than 2014 although 2014 is close behind. Also, Sep of 1987 is very close.

The comparison of the following two graphs of September SSTs 1948-2023 shows how much the equatorial region N of Australia has cooled in September of 2023 vs recent Septembers while at the same time Nino 4 directly to its E has warmed in September of 2023 vs recent Septembers. I assume this bodes well for increased chances of a colder winter for the E US vs recent ones assuming this relationship holds into winter: Much cooler equatorial area N of Australia Sep: Much warmer Nino 4 September:

Not surprisingly based on the animation, the anomalies for the equatorial W Pac area that I’ve been watching (120E to 160E, 15N to 15S) cooled even further from Aug to Sep allowing Sep of 23 to be a whopping 0.8C/0.7C/0.7C cooler than Sep of 22/21/20. This area had been very warm in recent years: https://psl.noaa.gov/cgi-bin/data/timeseries/timeseries.pl?ntype=1&var=SST&level=2000&lat1=15&lat2=-15&lon1=120&lon2=160&iseas=0&mon1=0&mon2=11&iarea=0&typeout=1&Submit=Create+Timeseries

Looking more closely: - During the first 7 weeks of your 11 week elapsed period of animation: the far E Nino portion cooled slightly while the remaining Nino areas warmed between 0.4 to 0.7. -But since then (subsequent 4 weeks): while the far E Nino continued to cool, the remaining Nino areas either remained steady (Nino 4) or cooled slightly (Nino 3 and 3.4). -So, since all Nino areas either cooled or remained steady during the last 4 weeks of your animation, it may make it seem like no Nino areas warmed during the 11 week interval. This is almost like a visual recency bias of sorts. And the far E portion (Nino 1+2) did actually cool during the 11 week interval.

During that time (week centered on 7/12 to week centered on 9/27), the indices’ anomalies went from: Nino 1+2: +3.4 to +2.8 Nino 3: +1.6 to +2.0 Nino 3.4: +1.1 to +1.5 Nino 4: +0.7 to +1.1 ———— Also, I can see in that animation how much the anomalies for the equatorial area of the WPac area that I follow (120E to 160E, 15N to 15S) cooled from July to Aug and how much the southern half of this area continued to cool from Aug to Sep. even while Nino 4 to its E kept warming. Much of the area between New Guinea and Australia has anomalies that cooled 1.25 to 1.75 C from the July 12th week to last week: —————— Weekly Nino anomalies: https://www.cpc.ncep.noaa.gov/data/indices/wksst9120.for

The Sep SOI came in at -13.9. How does this compare to the Sep of the 49 prior El Niño events dating back to 1876-7? -Of the 49, these 12 had more negative Sep SOIs: 2015, 1997, 1994, 1991, 1982, 1972, 1940, 1923, 1914, 1902, 1896, and 1877. So, 2023 comes in more negative than ~75% of the others. -Of those 12 with stronger Sep negatives, five were super strong El Niño events (2015, 1997, 1982, 1972, and 1877). These had a range of -14.1 (1997 and 1972) to -20.0 (1982). The super strong 1965’s Sep was at -13.5 while the super strong 1888’s Sep was only down to -8.8. - Regarding the other seven Septembers with a sub -13.9, four were strong El Niño events (1991, 1940, 1902, and 1896) with a range of -16.2 to -18.8. Seven strong Septembers (2009, 1987, 1957, 1930, 1925, 1918, and 1899) were less negative or even positive in one case (2009’s +3.6). - The remaining three were moderate El Niño events (1994, 1923, and 1914) with a range of -14.1 to -16.2. - Five of the 49 El Niño events came in with a +SOI in Sept: 2009 (+3.6/strong), 1979 (+1.7/weak), 1904 (+0.5/moderate), 1887 (+5.2/weak), and 1880 (+8.1/weak).

As of Oct 1, the -SOI streak reached 43 days. Although the streak may end at any point soon because it is so close to zero, I’m now leaning to it continuing through day 49 and then ending on Oct 8. Assuming the 49 days were to verify, how would that stack up to the longest streak of prior El Niño events back to 1994-5 (as far back as dailies are readily available)? Only the two super strong events of 2015-6 and 1997-8 would exceed 49 days. They both had max streaks well above that. But 2018-9, 2014-5, 2009-10, 2006-7, 2004-5, 2002-3, and 1994-5 didn’t. Of course, there’s always the chance that there will be a longer than 49 day streak later in this event: https://data.longpaddock.qld.gov.au/SeasonalClimateOutlook/SouthernOscillationIndex/SOIDataFiles/DailySOI1887-1989Base.txt

The Sept 100-180W OHC came in at +1.03, down slightly from last month’s +1.09 and well below the June peak of +1.40. This drop was the last thing I had earlier expected for Sep. (though I realized Sep could easily come in lower once I saw it fall way back) and the clock is ticking. Unless this does something extreme like the sometimes considered analog of 2009, when it suddenly warmed 0.99 from Sep to Nov to 1.75 after having fallen back from an initial high point in June of +1.13 to a temporary low point of only +0.76 (similar to the 2023 pattern); or like 1991, when after sputtering between April’s +0.80 and Sept’s +0.60 it suddenly warmed 0.81 to Oct’s +1.41 and then another 0.30 to Dec’s +1.71; it will not ever get to the high 1 area that I had earlier thought had a good chance to occur: https://www.cpc.ncep.noaa.gov/products/analysis_monitoring/ocean/index/heat_content_index.txt

Nino 4 at 29.87 is likely the warmest on record for 10/1 at least back to 1982. Also, when comparing to other Nino regions at the same time, it is very impressive! But considering GW, including how much warmer the globe has been this year, is it the most impressive for 4 for 10/1? Compare to 29.0 in 82, 29.1 of 86/87, 29.2 of 91, 29.6 of 15.

The August PDO was way down at -2.47. There have been 15 other sub -2 Aug PDOs since 1854 with only 1876 during El Niño. This is how the E US subsequent winters went including PDO: 1876 (only other Nino): winter cold mainly due to Dec, PDO -0.15 (rose 2.01, the third sharpest rise of the 15 from Aug to DJF), snowy NYC in DJ, major snowstorm Atlanta in D 1881: winter NN NE, warm SE, PDO -1.51 1883: winter cold NE, NN SE but very cold J, PDO -1.80, snowy NYC in D; wintry ATL in J 1892: winter cold, PDO -0.77 (rose 2.08, the 2nd sharpest rise of the 15 from Aug to DJF), heavy SN JF Boston/NYC, one of heaviest SN on record Atlanta in J 1909: winter cold, PDO -1.72, very cold/wintry ATL in D 1916: winter cool NE, NN SE, PDO -1.96, snowy NYC 1920: winter NN, PDO -1.11, heavy SN Boston F 1921: winter cool NE, mild SE, PDO -0.40, Knickerbocker SN NC/Mid Atlantic in J preceded by historic ZR SE coast 1933: winter very cold NE mainly due to Feb, NN SE, PDO -0.32 (rose 2.60, the sharpest rise of the 15 from Aug to DJF), heavy SN DFM NYC/Boston, major SN ATL in F 1950: winter mild NE, NN SE, PDO -1.26, wintry ATL D 1955: winter NN, PDO -2.25, very heavy SN M NYC/Boston 2010: winter cool NE, cold SE, PDO -1.77, very snowy winter Boston/NYC/ATL 2011: winter warm, PDO -1.87 2012: winter mild, PDO -1.28, very heavy SN FM Boston/NYC 2022: winter very warm, PDO -1.70 - 53% of the winters were cool to cold in the NE but only 27% were in the SE. - 27%/33% of winters were mild to warm in the NE/SE, including the last 3 - Active winters NE/SE when PDO rose sharply Monthly PDO back to 1854: https://www.ncei.noaa.gov/pub/data/cmb/ersst/v5/index/ersst.v5.pdo.dat **Edited due to typo as it’s Aug to DJF, not Oct to DJF, sharpest rises I’m referring to

Today’s GEFS based PNA is saying a daily peak of +2+ is likely near Oct 10th. The following Octs since 1950 peaked at +2+ with these peak dates and these full Oct PNAs: 1953: 10/6 Nino +1.91 1963: 10/12 Nino +0.69 1965: 10/5 Nino +1.14 1969: 10/8 Nino +0.35 1979: 10/1 Nino +1.53 1980: 10/4 +2.45 1984: 10/8 -0.63 1987: 10/3 Nino +0.53 2003: 10/6 +0.97 2008: 10/3 +0.86 2010: 10/2 +1.80 2011: 10/2 +0.63 2015: 10/10 Nino +1.78 - All 13 peaked within Oct 1-12 just like 2023’s prog. El Niños were the most favored ENSO for these through 1979. - All but one of these Octs had a monthly PNA of +0.53+. Seven of these 13 were within +0.53 to +1.14. All but one of the 13 cool to cold E US El Niño winters since 1950 were preceded by Octobers with a PNA of +0.16 to +1.14 fwiw. So, as one who’d prefer a cool to cold winter, I’ll be hoping for an Oct PNA within that range. - As @George001just reminded me though, the PDO was sub -2 in August, which itself doesn’t bode well if one is hoping for a neutral PDO or +PDO in DJF. More about that in my next post.