raindancewx

My 2020-21 Winter Weather Outlook: https://www.scribd.com/document/479516526/Winter-2020-21-Outlook

It's interesting...as hot and dry as it has been here, haven't really been many fires in New Mexico. I'm not optimistic for overall snow pack with the hot/dry Fall, mostly because I think we have a hot/dry Spring too. But I am fairly optimistic for winter itself. The strongest La Nina winters, the lowest-sea ice cold ENSO winters since 2007, and +NAO La Nina winters all support a cold West winter look in some fashion. Above: Lowest sea ice extent in cold-enso years since 2007. Below: Five strongest La Ninas (left), and La Ninas with a DJF NAO value of +0.4 or higher (right). The 80-160 La Nina Ace composite also looks like these maps, and the Atlantic is at 122 ACE so far.

My outlook will be out tomorrow. But I was looking at something today: Since 1950, by my definition there are 24 winters following an El Nino winter. In those winters, if the NAO is positive in October, it is only positive in winter 5/13 times. But, if the NAO is negative in October, it is positive in winter 8/11 times. These are tiny samples, but if you do a difference in proportions test, it still just barely passes as statistically significant (p<=0.05), i.e. probably not a fluke or due to chance, and a real relationship. It's a shift in odds by ~35%. My NAO forecast blend already expected the NAO to be positive, but I was surprised to see any kind of correlation for this. It is pretty linear too: so if the NAO is -1.5 or something this October, you'd expect the NAO to be more positive in winter than if it were -0.5. The +NAO La Nina composite is this: The strongest La Ninas since 1950 (<25.0C in DJF) give this - don't really expect the La Nina to last at that strength for DJF though. The low-sea, cold ENSO years since 2007 are similar too. The 80-160 La Nina ACE years (currently 122) also look a lot like the first two composites.

I like 1954 as an analog. My statistical predictors for temperatures in the Southwest have that year as the best match to 2020 - huge number of hot Summer days, similar ENSO, ENSO trends, and solar, weak monsoon in the US, India and East Asia. For La Nina years the more recent events tend to be surrounded by warm waters in the Pacific that were not really present in older years like 1954. I view La Nina winters in the US as more Atlantic driven than Pacific driven in most spots, so the mid-1950s to 1963 representing the late warm AMO of that era is similar too, Where I live, the top matches for June-September in a La Nina are 2007, 1973, and 1954. My guess is the similarities to 1954 and 2020 have to do with relatively early substantial La Nina strength late Summer in both years and similar MJO trends and similar net ACE output in the Atlantic. But the MJO data doesn't exist before the 1970s. The sea ice extent in some years in the 1950s probably dropped to around 5 million square kilometers in September, but that is still well above this year too, which has to have some effect. Where I live, the coldest La Nina years tend to follow a) early heat (first 90 degree Fahrenheit (F) years), b) consistent heat - I know 1954 had 118 days 87F or hotter from April-October. We're at 111 this year. Both had 90 degree heat by mid-May If I did a purely ENSO based winter forecast, I'd probably blend 1954, 1964, 2007 for the winter, double weighting 1954 and 2007. For an American perspective, one of the things I like about 1954 is it is a La Nina with an El Nino after a very warm El Nino winter as 2020 will be. Something like this is probably the best case scenario for the West. It seems to be favored look in low-sea ice cold ENSO since 2007, in cold-ENSO after El Nino (1931, 1954, 1988, 2007, 2016), and in strong La Ninas (2007, 1988, 1954).

Cold ENSO with low sea ice since 2007...v. Cold ENSO with (relatively) high sea ice since 2007. Really curious to see if this works in 2020. I don't really look at 2007, 2012, and 2016 as similar...they are very different ENSO events. But they all have the low-sea ice and cold West. I put 2011 in both because it was cold in NM/TX at the border of the two cold areas, and the 4.3m square km also seems to be the borderline threshold for this. The ACE index is higher on the left image, so maybe this is just all an Atlantic thing, since both sea ice and ACE are tied broadly to the AMO.

I'll go into it more later, but I like 2003, 2012, and 2019 as "fixers" more than anything. Although 2003 is pretty close to the temperature pattern for this October on the models. It actually is one of the better matches on some non-ENSO stuff I like. In general, 1995, 2003, 2007, 2019 all saw declining Nino 3.4 values year over year. 2012 is an interesting case, as it is around 27.0C in Nino 3.4 late Summer but fell off hard to 26.0C by January - roughly - I'm approximating. I actually Googled "Paul Pastelok Winter Analogs 2020-21" just to see if anyone else had similar years, and he had 1995, 1998, 2003, 2005, 2010, 2016 as his package for winter as of August (may have changed since - not sure) - and if you look at the Accuweather winter forecast it does kind of look like that blend. I wouldn't pick 2010 or 1998 - you're talking about a massive year/year change in SSTs both times that we just don't have in Nino 3.4. 29C to 25C or whatever 1998 was...is nuts. This is 27C to 25C best case, and probably 20% less than that. I think 29C to 25C is close to the maximum downward change you can get year/year in Nino 3.4. My blend has super low ice. The ice got to 3.74 million square km in September 2020, the main analogs, 2007/2012, are pretty dead on to that. I've mentioned before, the lowest sea ice extent years with a cool ENSO event, starting 2007, have tended to produce similar cool West looks. 2007, 2012, 2016 fit this. Not a higher sea ice year since 2007 does: 2008, 2010, 2013, 2017, with 2011 right at the edge (the line seems to be 4.3m, and it was cold in NM/TX). Weatherbell currently has 1973, 1988, 2010, 2010 as their blend. That'd be a decent winter here, but I don't think it's right. Nino 4 cooled way faster in 1973 and 2010 than in this year. 1988 may not be terrible for the "look" of winter, but its probably a much stronger La Nina than this year, with very high solar, and a colder North Atlantic and North Pacific. September 2019 (very warm) and September 1973 and 2010 (very cold) are all about 1.0C out from current readings in Nino 4. I like 2007 because Nino 4 did cool off a lot, but it took much longer than 1973 and 2010.

This is the Euro plume from October 1 for the La Nina. I've drawn in my expectations - I think this event could peak pretty strong. But I also think the Euro is right about a pretty quick turnaround/warm up after January.

This is what I ended up with for snow nationally. The numbers are percentages of normal for July-June. My forecast is basically done, just waiting for a few things to come in this week.

I mean...I've seen 90 point dew point depressions here. You all are just spoiled up there. The -PDO can really dry things out in the West if it lasts long enough, and Nino 4 finally getting knocked below average helps too. The dew points here tend to run in the 30s/low 40s into October in a good monsoon year most days. When the monsoon fails it seems like dew points run from 10-30 most of the time in Fall. That's one reason I don't think it's a super hot winter though. It's going to be pretty dry, but with semi-regular cold fronts I think. So cold air moving in when dew points are -10, 0, 10, 20, etc. is a recipe for some pretty cold nights. I've been looking at precipitation patterns by hurricane season activity here. December has a near 0 chance of topping average precipitation in a very active hurricane season (>160 ACE) in a La Nina. But it has a 50% chance in the other La Ninas, going back to the 1930s. The difference is statistically significant (p is 0.037 or so), and none of the other cold months are even close to that. ACE is still only at 111 despite 25 named systems. So December has a decent shot at being a good month. Not really optimistic about November even though I kind of expect a big cold dump in late October, probably the week before Halloween. I kind of get the impression there will be a good storm or storms by the end of October for much of NM, CO, WY.

ACE on the CSU site has updated at least twice today and is currently 111.4 as of late Tuesday. It seems generally you add 1 point for a weak tropical storm that lasts all day, 2 for a weak hurricane that lasts all day, 3 for a major hurricane, and 4 for a category 5. This system has four days left until it dies, so the ACE should get into the low 120s, which is still the mid point of the 80-160 La Nina composite favoring a cold NW. By 10/10, the season is typically 85% or more over, but it can be less, and it can be more. I lean toward a couple more tropical storms after Delta, but this is probably the last big system this year. You'll probably see some big weakening if Delta hit Mexico, and then a larger, weaker hurricane after. The Gulf of Mexico is pretty spent at the moment, so it is fairly cold. So I lean toward 2.5 points a day for the next four days from Delta. We'll see how close 10/10, at 5 pm MST is to 121 ACE. I'm planning to release my outlook 10/10. I've got 50+ slides this year, but most of it is pictures. I've got a pretty hot winter in some parts of the South - mostly the coastal SE and especially Texas. I have the interior West cold, but not that cold. I have most of the East warm, but still much colder than last year. Statistically, the odds of a wet December here are near 0% if the ACE is above 160, and near 50% if the ACE is under 160 in La Ninas over the past 90 years. So do want to see how high the ACE gets. Since mid-September my outlook has assumed years with 95-175 Atlantic ACE would be similar, given about 135 ACE this year.

As best I can tell, there are only 17 hurricanes to hit the US Gulf Coast in October since 1931. It's interesting to see 1964 show up as one of the years (Hilda), with one of the once a decade non-Florida hurricane landfalls on the Gulf in October. Have to see if Delta actually follows the current projected track. The closest tracks since 1931 in October to Delta are probably Nate (2017), Hilda (1964), Opal (1995), Lili (2002), and Jerry (1989). Of those years, 1964 is the only one with a warm NW October though (currently 5-10 above normal in some places in Oregon).

This is going to be a really weird winter for our region. Pretty sure of that. The storms are finally starting to show up in the long-range for the Northwest, which is good, that was what I expected. My best guess is the NAO will go negative, either substantially but only for a small time, or only a bit, but for an extended period, in December. I've had some good success using this method in prior years. March would likely be pretty negative too. My sense is the intense heat of Summer will show at times in winter, but you'll also see very intense waves of cold show. The worst case scenario for us would be that my reliance on ENSO order is irrelevant, and years like 2008, 1999, 1975, 2017 end up taking over somehow, instead of the La Ninas that follow El Ninos like 2016, 2010, and 2007 - relatively wet years generally.

General warming again in Nino 1.2 Way ahead of 2016 and 2017, 2008 at this point. Somewhat colder than 2011 too most spots. Nino1+2 Nino3 Nino34 Nino4 Week SST SSTA SST SSTA SST SSTA SST SSTA 02SEP2020 18.9-1.6 23.6-1.3 25.8-0.9 28.4-0.3 09SEP2020 19.5-1.0 23.4-1.4 25.7-1.0 28.5-0.2 16SEP2020 20.0-0.4 23.6-1.3 25.9-0.8 28.2-0.4 23SEP2020 19.6-0.8 23.6-1.2 25.7-1.0 28.1-0.5 30SEP2020 20.1-0.5 23.8-1.1 25.6-1.1 27.9-0.7 30AUG2017 20.3-0.2 24.5-0.4 26.5-0.2 28.8 0.2 06SEP2017 20.4-0.1 24.3-0.6 26.2-0.6 28.7 0.1 13SEP2017 19.7-0.7 24.0-0.9 26.1-0.6 28.7 0.0 20SEP2017 19.3-1.1 23.9-1.0 26.3-0.4 28.7 0.0 27SEP2017 19.5-1.0 24.4-0.5 26.5-0.2 28.4-0.2 31AUG2016 20.9 0.3 24.6-0.3 26.0-0.7 28.5-0.1 07SEP2016 20.8 0.3 24.6-0.3 26.1-0.7 28.6-0.1 14SEP2016 20.6 0.2 24.7-0.2 26.1-0.6 28.5-0.2 21SEP2016 21.2 0.8 24.8-0.1 26.3-0.4 28.5-0.2 28SEP2016 21.2 0.6 24.7-0.2 25.9-0.8 28.1-0.5 31AUG2011 20.0-0.5 24.4-0.5 26.1-0.7 28.3-0.4 07SEP2011 19.7-0.7 24.2-0.7 26.1-0.7 28.2-0.5 14SEP2011 19.6-0.8 24.1-0.7 26.0-0.8 28.0-0.7 21SEP2011 19.8-0.6 24.3-0.6 25.9-0.9 27.8-0.9 28SEP2011 19.9-0.6 24.1-0.7 26.0-0.7 28.0-0.7 03SEP2008 21.3 0.8 25.3 0.4 26.6-0.2 28.1-0.6 10SEP2008 21.0 0.6 25.2 0.3 26.6-0.2 28.1-0.6 17SEP2008 21.1 0.8 24.9 0.1 26.4-0.4 28.1-0.6 24SEP2008 21.3 0.8 25.1 0.2 26.4-0.4 28.1-0.6 01OCT2008 21.0 0.4 24.7-0.2 26.2-0.5 28.3-0.4 Still way warmer than 2010 though in all zones. 01SEP2010 18.8-1.7 23.4-1.5 25.1-1.7 27.2-1.5 08SEP2010 19.1-1.3 23.5-1.4 25.2-1.6 27.1-1.6 15SEP2010 18.6-1.8 23.4-1.5 25.1-1.6 27.1-1.6 22SEP2010 19.3-1.2 24.0-0.8 25.1-1.6 27.1-1.6 29SEP2010 19.2-1.4 23.5-1.3 24.8-1.9 27.1-1.6 06OCT2010 18.7-1.9 23.2-1.7 24.8-1.9 27.1-1.6 Very close to 2007 in Nino 3.4 and identical in 4. Pretty different in Nino 1.2 - but 2007 warmed up a lot there in December, so might snap back toward 2007 then. 1995 is close too. 05SEP2007 19.4-1.1 23.7-1.2 26.0-0.8 28.3-0.4 12SEP2007 18.6-1.8 23.6-1.3 25.8-1.0 28.0-0.6 19SEP2007 18.5-2.0 23.6-1.2 25.8-0.9 28.0-0.7 26SEP2007 18.4-2.1 23.2-1.7 25.6-1.1 28.0-0.7 03OCT2007 18.3-2.3 23.2-1.7 25.4-1.3 27.9-0.7 06SEP1995 20.4 0.0 24.1-0.8 26.0-0.8 28.2-0.5 13SEP1995 19.9-0.5 24.1-0.8 26.0-0.7 28.0-0.6 20SEP1995 20.2-0.2 24.0-0.9 26.1-0.7 28.4-0.3 27SEP1995 20.2-0.3 23.9-1.0 25.9-0.8 28.4-0.3 30SEP2020 20.1-0.5 23.8-1.1 25.6-1.1 27.9-0.7 The September NAO was more positive than I expected. When I run my NAO script, it suggested several possible blends, but I think this one is probably the best for what it will do: 2020: 1961, 2003, 2007. March would also be somewhat negative in this scenario. In prior years, 1993, 2017, 2017 was a good blend for 2019. Generally, you look for the best match that has similar ENSO tendencies for May-Apr and Sept-Mar for the NAO state. If you don't have a good ENSO match available because of some unusual combination of NAO changes, the best blend overall will typically work, based on testing from 1995-2019. 1975 is a good match to 2018 1990, 1990, 1994 was a good match for 2017 2004, 2010, 2011 was a good match for 2016 1992, 1994 was a good match for 2015.

All of the Nino zones came in below the 1951-2010 historical averages in September. Compared to September 2007, the current La Nina is consistently 0.2-0.4C warmer in all zones Nino 4 Nino 3.4 Nino 3 Nino 1.2 2007 28.11 25.61 23.63 19.75 2020 28.28 25.96 23.93 19.54 In the 1950-2019 data, you don't have any years where Nino 3.4 dropped off by more than ~1.0C from September to Dec-Feb. So it's pretty likely this event will come in 25.0C or warmer, as the subsurface data had previously implied. I think a couple months could come in around 25.5C - but we'll have to see.

Year Jan Feb Mar Apr May Jun Jul Aug Sep DJF 3.4 DJFp 3.4 1966 -21.90 -17.14 -11.07 -2.33 2.16 5.42 7.47 7.63 9.23 26.18 27.78 2010 -16.02 -16.98 -19.68 -23.57 -26.28 -25.05 -9.84 1.45 6.58 25.21 28.14 2016 9.34 6.77 3.16 0.64 2.37 3.86 6.25 10.07 10.48 26.30 29.13 2016 9.34 6.77 3.16 0.64 2.37 3.86 6.25 10.07 10.48 26.30 29.13 Blend -4.81 -5.15 -6.11 -6.16 -4.85 -2.98 2.53 7.31 9.19 26.00 28.54 2020 -2.51 -3.20 -4.36 -5.03 -4.86 -2.78 0.34 4.78 7.95 Trend down down down up up up up up That's the simplest good QBO match I can come up with using the September data, although the drop off in Nino 3.4 won't be ~2.5C. Probably more like 1.8C. If the QBO trend matters, 2008 may be peaking as a similar year now as the QBO dipped from September to November. I think the raw value for 2020 could be ~15 in November as it peaks or continues to rise, while 2008 fell off to 9 or so from a June/July peak before a second winter peak. I'd imagine the current QBO peak will not be a double peak like that. Throwing in 1961 would keep the trends/match pretty strong but make the y/y more comparable to this year. Either way, still kind of a cold NW / warm East & South look nationally. Do want to see the monthly SST data for September. I think its pretty close to 2007 and 1995 still. The MJO lingering in phase five around 10/1 in a low solar La Nina is something I take as a good sign for what I expect in winter as it happened in 2007, 2011, and 2016 to some extent. The blend of those years looks a lot like what I have, and gets you a ~25.5C La Nina winter, following a 27.2C El Nino the prior year (27.1 in 2019-20). Similar ACE (104 v. current 105), low sea ice, with a -PDO, but not out control. I do think the -PDO looks less likely to be severely negative than a few weeks ago, but probably still negative. But I lean more toward 0.0 to -1.0 for Nov-Apr on the Nate Mantua method than the -0.5 to -1.5 I was expecting before. If the entire North Pacific is cold or warm, you can still have a + or - PDO if the relational setup between the NE & NW are correct, so I lean toward a "warm" version of the -PDO for Nov-Apr. Also: I'm not optimistic for the NE overall, I think it will be warm again, but I actually do think it's a pretty decent snow setup for New England. I'm not talking 80 inches of snow for Boston, but a general +0% to +30% for snow. I'd have to look, but I think NYC was near average, Boston was above, and south of NYC was pretty crap in the blend I have. The low solar La Ninas are pretty good for Boston for snow in the past 90 years: 1933, 1942, 1954, 1964, 1973, 1974, 1975, 1984, 1995, 2005, 2007, 2010, 2016, 2017. Most of the years are above average (9/14) and no true duds. Even without 1995-96, the average is 46 inches, so the 50 inches overall isn't super inflated. 1933: 63" 1942: 46" 1954: 25" 1964: 50" 1973: 37" 1974: 28" 1975: 47" 1984: 27" 1995: 108" 2005: 40" 2007: 51" 2010: 81" 2016: 48" 2017: 60"

The cold shots have been showing here every 45-days (ish) since July, arguably May here, despite a very warm overall pattern since April. That would put the MJO induced repeat of the September snow storm around the week before Halloween. I do think October will be warm overall, but the pattern looks like it will break down. Long-term, I'm not optimistic for snow pack overall, mainly due to the shoulder seasons being warm/dry in general. I don't think the actual winter will be that hot or dry for Colorado, or even northern New Mexico. The current La Nina is probably the 3rd to 5th strongest to develop since 1980 at this point. It's behind 2010 and 1988, kind of tied with 1999, 2007, and 2011 by the metrics I'd use to estimate strength. A lot of those winters aren't actually that bad: 1988, 2007, 2010, 2011 are all OK to good down here. The weaker La Ninas tend to be hot/wet, or bone dry, with average temperatures. The stronger ones feature very few real storms, but often a lot of moisture starved cold fronts where you can sneak out 0.05" as snow each time, and then enjoy the cold wave for 2-5 days, or even 5-15 days in some cases.

There seem to be two areas impacted by the QBO in the US from what I can see. For whatever reason, there is a pretty big difference in the South Central US between La Ninas with a QBO -2 or lower and La Ninas with a QBO +2 or higher. The same NW areas are still cold though. The years with a drop off in QBO values in Summer (JJA) to Winter (DJF) seem to feature at least some cold dumps into the central US more reliably. So that's the coldest La Nina/QBO composite I could find. The rising QBO years are more amplified, with broader warmth and cold in the NW. We aren't in the first situation. The QBO was about 0 in JJA and rising. That map (rising) as a composite gives the same features as winters with a QBO >2. So unless you live in Kansas, most areas are pretty similar regardless of what you do with the QBO. These composites are so small and erratic, that I'd say they mean almost nothing in aggregate. I did draw in some dummy high/low positions without checking the maps as a guess for how the composites might be varying. I'm personally much more interested in the ACE value for the Atlantic at this point. The long-term average is about 105.6 so we're still technically below an average season (1981-2010), and La Ninas average 128 ACE from 1930-2019. October averages about 15 ACE, and I doubt there has ever been an October ACE over 50. Locally, there is a very strong correlation in La Nina years to how cold we get at the coldest time of the year here - Dec 16-Jan 15, when the average high is 46 and the average low is 24. So quite curious to see how high the ACE gets. The projected mid-Dec to mid-Jan value high on the image below is reliably (~5/6) within 3.3 degrees of observations, and actually tends to be a lot closer than that if ACE isn't under 65 or so where most of the outliers are. I've looked at Neutral and El Nino years here, and you don't have a relationship like this in any month from October-May. The relationship holds well over time too - 1933 is the dot furthest to the right with the highest ACE. My sense is a low ACE La Nina went into phases 4-5-6 fairly often during the heat of the hurricane season. Five, or the "five look" with a warm Nino 4 can be very cold here if it shows up again in late December like in 2007. Below the surface, 100-180W looks a lot like a combination of 2003, 2007, 2010, 2017 for July-September, and the match is fairly close to what the models show for October and how September ended up. I'm not a fan of 2010 and 2017, they both have timing and severity issues. There are non-ENSO things I like about 2003, but can go more into that when I put my outlook up in a week or two. https://www.cpc.ncep.noaa.gov/products/analysis_monitoring/ocean/index/heat_content_index.txt

Canadian has a long lasting La Nina that gradually becomes centered pretty far West. It trended stronger than before. Nino 4 would stay cold for a long time in this depiction - that hasn't happened in close to a decade. The monthly sea surface temperature readings for the September monthlies should look a lot like 2007. On the Canadian, the cold ring by NW North America tries to develop around the warm tongue east of Japan. That'd be a pretty healthy and canonical -PDO, which is a pretty warm signal for the SE. Good for the NW to be cold. Pretty strong dry signal in the SW, especially Fall and Spring, locally it peaks as an indicator in November and May.

Still looks a lot like 2007, especially in Nino 3.4 and 4. Nino1+2 Nino3 Nino34 Nino4 Week SST SSTA SST SSTA SST SSTA SST SSTA 05SEP2007 19.4-1.1 23.7-1.2 26.0-0.8 28.3-0.4 12SEP2007 18.6-1.8 23.6-1.3 25.8-1.0 28.0-0.6 19SEP2007 18.5-2.0 23.6-1.2 25.8-0.9 28.0-0.7 26SEP2007 18.4-2.1 23.2-1.7 25.6-1.1 28.0-0.7 02SEP2020 18.9-1.6 23.6-1.3 25.8-0.9 28.4-0.3 09SEP2020 19.5-1.0 23.4-1.4 25.7-1.0 28.5-0.2 16SEP2020 20.0-0.4 23.6-1.3 25.9-0.8 28.2-0.4 23SEP2020 19.6-0.8 23.6-1.2 25.7-1.0 28.1-0.5 October is depicted like a relatively canonical MJO phase five October, and an almost identical match to 1964 on the CFS

September looks like it will finish near average for temperatures here. One of the driest monsoons since 1931 for Albuquerque. The current October forecast on the CFS is very similar to 1964. Relatively canonical MJO phase five look for October, which makes sense as October is expected to be in phase five initially.

Nino 4 looks like it is cooling at the surface, but still pretty warm below. It does look like the cold is bleeding West, which is what the models show, and what happened in years like 2007.

There is another nice cool down coming in a few days after some unusual late heat. CFS currently has a super hot October for much of the Northwest, but I have my doubts. Will likely change meaningfully in the next few days.

I don't do Fall outlooks, but if I had done a Fall outlook, you really can't find a better match long-term than 1995 for the Nino zones. But I do expect 2007 to take over later in the fall. This is Nino 3.4 1995 27.57 27.49 27.75 28.10 27.82 27.59 27.08 26.23 2007 27.24 26.88 27.10 27.50 27.46 27.37 26.71 26.14 2020 27.14 27.11 27.76 28.17 27.65 27.38 26.99 26.30 Nino 4 - 1995 29.04 28.91 29.03 29.14 29.31 29.18 28.93 28.65 2007 28.97 28.66 28.66 28.73 28.84 28.90 28.68 28.48 2020 29.16 28.97 29.07 29.15 29.00 29.09 28.89 28.50 The biggest problem with 1995 is the PDO I think - 2020-03-01T00:00:00Z -0.82 2020-04-01T00:00:00Z -0.57 2020-05-01T00:00:00Z 0.09 2020-06-01T00:00:00Z -0.08 2020-07-01T00:00:00Z -0.38 2020-08-01T00:00:00Z -0.28 https://oceanview.pfeg.noaa.gov/erddap/tabledap/cciea_OC_PDO.htmlTable?time,PDO 1995-03-01T00:00:00Z 0.75 1995-04-01T00:00:00Z 0.83 1995-05-01T00:00:00Z 1.46 1995-06-01T00:00:00Z 1.27 1995-07-01T00:00:00Z 1.71 1995-08-01T00:00:00Z 0.21 90% of the time, maybe more, the PDO values are -2 to +2. So +1.7 in July 1995 and -0.4 in July 2020 is a massive difference. Even the August difference is still pretty substantial. It popped back up too - to +1.2 in September, and on/off again was pretty positive through Spring. Nino 1.2 has been warming. If that continues into October, then the PDO won't go as negative as it looked a few weeks ago. If it really warmed up a lot, it might even go positive, but I don't expect that. You can see cooling lately in some of the areas you'd want for a -PDO on Tropical Tidbits (by the coast of Western Canada), but also hints of cooling east of Japan where you'd want cooling for the +PDO. This is what I have tentatively for winter, with the analogs removed, since people need to come up with their own ideas - Here is how the analog package looks compared to September. It is also similar on solar, ENSO trends, and actual US temperatures for July-September (assuming the forecast NE/Western heat continues to burn off cold in those areas). You can see that the transition in the La Nina is the same in my analogs as on the CFS. I was a bit surprised to see the cold eastern Atlantic show up in September in both 2020 and the analogs, because it definitely warms up by winter.

The analogs I have look like the -QBO years because 2007 is the analog I trust most, even though it is still flawed. However, late winter and spring do look like the +QBO years on my blend for the winter. 2007-08 had a -QBO but it was rising, like this year. I find that the trend is about as reliable as the actual QBO anomaly for large areas of the country. I'm skeptical to bet the farm on only ENSO/QBO DJF seasonal averages. To me 2010-11, 1995-96, 1988-89, 1973-74, 1964-65 are also very different winters. I actually don't consider 1995 a bad analog, but it's mostly for how similar the ENSO has been all year, not the surrounding stuff like the PDO or AMO. It's not going to be similar on hurricane activity either, we're at less than half the 1995 total ACE in the Atlantic to date even with 23 named systems. Here is how these years looked for the QBO v. 2020, to me, they look very different with unusual tendencies in 2020 that would suggest erratic QBO behavior later on. I don't have QBO data at 45mb, but this is 30mb. 1964 3.94 5.26 5.46 5.94 6.32 2.23 -0.56 -0.57 -0.67 0.42 0.71 0.04 1973 -4.40 0.08 3.40 6.28 8.12 8.63 6.94 5.86 5.51 5.20 4.92 2.31 1988 7.81 6.17 5.86 6.59 5.46 0.42 -3.96 -2.58 -2.29 -1.53 -0.84 -2.42 1995 8.38 8.01 8.79 11.79 14.92 15.62 11.74 9.53 6.98 3.43 -0.77 -4.57 2010 -16.02 -16.98 -19.68 -23.57 -26.28 -25.05 -9.84 1.45 6.58 10.83 12.16 10.97 2020 -2.51 -3.20 -4.36 -5.03 -4.86 -2.78 0.34 4.78 -999.00 -999.00 -999.00 -999.00 64 up up up up down down down down 73 up up up up up down down down 88 down down up down down down up up 95 down up up up up down down down 10 down down down down up up up up 2010 down down down up up up up up You can see trend wise, 1964 is backwards (opposite) as is 1973. 1995 trended down mid-year, while we are trending up. You do have the right idea for Jan-Aug in 2010, but the QBO looks to me like it will keep going up the rest of the year. It didn't in 2010. Does that matter? Maybe? The magnitudes seem like they could be pretty close in 2010 through Fall and then be very different in winter. If you care about the trend I think 2016 is probably better than 2010 as a match. At a high level, I know 2010 and 1988 were both active monsoon years, without record setting heat in the West in the Summer. 1995 is closer in that regard too. To me, if the QBO was meaningfully similar through August, you wouldn't see something like a near all-time strong Western Monsoon in 1988 and a near all-time weak Western Monsoon in 2020. Some of that is solar driven, but the QBO is supposed to impact tropical weather and by extension subtropical weather like the monsoon. I think part of it is Nino 4 dropped much faster in 2010 and 1988 than this year. It had dropped like from 29.5 to 27.5 or something in 2010 January to August. It's only fallen 0.6 or so this year. My real issue with 2010 is you're talking about a ~3C year/year drop in Nino 3.4 We're not getting that this year, this isn't the La Nina version of 2015-16. Even 1988 had like a 2.5C drop. I'm looking for like a 1.6C drop y/y (27.1C to 25.5C).If you pulled up August 1964, it was super cold in the entire West - complete opposite this year. 1973 is warmer, but still nothing like this year. In 1995, you have a similar looking August and September nationally, with the slower decay of Nino 4 like in 2007 and this year.

This is the updated subsurface comparison to 2007. To me it kind of looks like 2007 but everything happens 4-8 weeks later. So maybe the September flat-line in 2020 is the 2007 flat-line in August, and then the La Nina peaks in October? I guess we'll see. It's possible though that we really won't see much change in the subsurface for a while.