I posted this the second week of September on a couple of forums, just posting it here for the record/reference later…
Playing with some analogs right now…here are some analogs, discussion, and caveats, focused mainly on the U.S….FWIW, my temp map for work is not as cold as this analog composite would imply and allows some risk for a SE ridge to crop up should things trend too far in the wrong direction (read below).
Here are my key assumptions when looking for analogs right now…the analogs used to compile the attached 500mb map were found by subjectively rating every possible analog on each of the following criteria. Ultimately I give ENSO, PDO and QBO the most weight along with the solar cycle, though all of these are considered.
- ENSO: Neutral, though can allow for weak El Nino or weak La Nina if the year featured warmer waters near the dateline. Extra preference if coming off of a weak or moderate El Nino the prior winter-spring.
- PDO: Neutral or positive (it’s positive right now though not strongly)
- QBO: Positive trending negative, expecting the 30mb winds to flip to negative at some point in the early to mid-winter
- Solar: Minimum
- Indian Ocean: Positive IOD in the fall, can trend downwards during the winter
- Off-equator Pacific SSTs: A positive to strong positive PMM (it’s very positive right now), with more weight given if it stays positive through winter
- Atlantic: Neutral or positive AMO (it is positive right now though not strongly)
- Tried to find matches based on similar tropical forcing in mid-late summer as this year, though not as strong of a weight.
- Tried to find matches based on years with significant spring/summer high-latitude blocking, though was not as strong of a weight.
The analogs I went with for this exercise are…1958-59, 1966-67, 1969-70, 1978-79, 1980-81, 1985-86, 1990-91, 1993-94, 1995-96*, 2003-04, 2004-05*, 2013-14*, 2014-15, 2017-18. Stars denote highest scoring matches when considering all of the above and double weighting.
Obviously some of these are very cold winters for the eastern U.S. and the composite look is cold. There is a strong signal for Alaskan ridging and a -EPO, which is not a warm pattern for the central and eastern U.S., though where exactly the cold dives in can make a difference for the eastern U.S. with more mixed signals on an NAO. A neutral-ish ENSO and +QBO to start winter aren’t great signals for a -NAO, though the deep solar min is and the QBO will be improving through the winter. The SSTs up there support a -NAO, but aren’t a strong forcing mechanism on their own. There is not a strong correlation between negative summertime NAO and subsequent winter NAO and the forcing mechanisms are different, so the persistent -NAO this summer doesn’t really help or hurt.
In terms of what to watch for in the eastern U.S. in terms of swinging warmer or locking in cold…I’ll be watching to see if we hang on to the warm waters near the Dateline in the Equitorial Pacific and the +IOD as we head into fall. If we keep those we are more likely to see convection near the Dateline this winter which usually forces an Alaskan/western Canadian ridge. If we see a stronger push towards La Nina and lose the warmth near the Dateline, the risk for a more amped SE U.S. ridge increases…we don’t need the Nino region 3.4 anomaly to be above 0C to have a cold winter in the east, but region 4 is pretty important and needs to stay warmer IMO.
Based on the persistent -SOI and forecast generally weak trades over the central and western Pacific over the next week or so, there won’t be a big La Nina push in the near-term, though the recent easterly trade surge did nudge things in that direction over the last couple of weeks and there’s still a lot of time for that to resume. The waters near the Dateline and just west remain fairly warm both at and below the surface, and until that goes away some move back towards a weak Modoki El Nino also can’t be ruled out. As we saw last winter, a SSW can really enhance tropical forcing/convection over the West Pac warm pool (which is usually warm for the eastern U.S.), so an initially +QBO and seemingly low risk for an early SSW may give some margin for error…but if the SST pattern becomes unfavorable between Australia and S. America for convection near the Dateline the pattern more likely supports eastern U.S. warmth this winter.
The analogs that have an Aleutian low in October generally had much colder subsequent winters than the ones that have an Aleutian high and subsequent trough over western Canada in October. I’m aware of what the longer range guidance hints at to start October up there, but wouldn’t lock it in yet. A continued drop of the QBO heading into the fall is also important for increased high-latitude blocking prospects as we head into winter and a lower risk for the Pacific jet to be too strong/zonal into the west coast, which would likely result in quite a bit of warmth for North America given how mild the entire Pacific is.
An additional caveat is the northern hemisphere water temperatures from the tropics up to the polar regions where there was another near-record sea ice melt are ON FIRE. Do older analogs break down as we continue to warm? If so, how do things change? The warming baseline gives less margin for error (you probably don’t luck into below-average or even average temps, you either have cold signals or you torch).
Lots of food for thought…I definitely don’t hate the prospects for a snowy winter from the Ohio Valley and Great Lakes into the Northeast, and the warm PMM and any remaining warm ENSO influence can bring some snow prospects to the mountains out west, but the pattern can turn warm quickly if we lose the warm equatorial waters closer to the Dateline.