raindancewx

August SOI is traditionally a much stronger signal for a warm Northern US December than Nino 3/3.4/4. Month looks like it will finish among the lowest ever SOI readings for August. Vast area of Southern Canada/Northern US favored warm at a strong correlation already. For whatever reason, the August SOI is more predictive for Dec temps than the Sept SOI as well.

Southern South America has been seeing a cold winter (their winter is our Summer). That's fairly normal for an El Nino "expected effect". The Southern Hemisphere +PDO is sort of there, warm ring around a cold tongue. Areas expected to be warm in El Nino have been running +10F or more in winter - good preview for our winter in the northern US I'd expect.

The recent El Nino Summers (2015, 2018, 2019, 2023) as a blend are pretty close to 2026 for Jun-Aug - not quite by magnitude but in terms of pattern. Close match for Sept on the CFS too. Honestly, it's probably a decent match overall. 2015/2023 - very strong El Ninos with high solar, neutral/negative PDO values in all four years. Warm Atlantic. 2015/2018 both followed very warm/dry Western winters like 2026 will. Decent QBO match too.

The Canadian outlook for August verified way too cold for the US. Far too wet too. Worth keeping in mind for what it shows over the next few months. Fall months are often the hardest to predict anyway, particularly October, so not expecting much accuracy from these outlooks for a while. Colorado is +5F in areas expected to be -1F to -2F v. long-term averages as an example.

I've been disappointed with the E Pacific hurricane season so far. Was hoping we'd have several land falling systems on Western Mexico to help with rain out here. Nothing. CFS is insistent every month from Sept-Mar will be wet locally. We could use that. CFS also has the entire US, except for New England, cold at least one month from Jan-Mar, so that's encouraging to see. For comparison, 2023 had 11 named systems in the Atlantic by 8/31. We're at...four. By the end of August, 2015/2019 had five, 2018 had six, and 2014 had three. 1982/1997 were relatively quiet too. 2004/2018/2023 are the more recent active hurricane seasons in El Ninos. Nice to see we're not getting a big hurricane season with this. Wyoming is favored wet in El Nino Augusts. Other than that, precipitation is still not behaving with an El Nino look. No AO/NAO/WPO/EPO/PNA signature to it either.

Going to be very interesting to see if we get 'normal' El Nino conditions here. So far the Summer here has been about as hot as ever recorded, with near record low precipitation as well. 2023 was even drier in the Summer, but we're running hotter. Summer has been very close to 1953, 1958, 2023 as a blend, all the sort of drought plagued, -PDO, El Nino, very +AMO shit years locally. It's actually been very interesting - for the past month little batches of rain showers have formed late evening almost every night. But they are widely scattered and have only brought drizzle to the city most nights. A lot of the terrible monsoon years don't even have the showers form. But they typically perform best for rainfall late afternoon to about 8 pm. We've seen showers develop in the 8 pm to 2 am time frame and then sizzle out, which is unusual here for formation time. The Atlantic has really warmed up quite a bit relative to averages over the past 2-3 weeks. I think we will actually see a brief period of decent hurricane season activity in September. My hottest average temperature El Nino Summers are these winters: 1994, 2023, (2003 borderline El Nino), 1958, 2018, 1953, 2002. We tend to get fluky out-of-season heavy snow events following very hot and very dry Summers here. 1994-95 snowed a lot in April, 1953-54 in November, 2002-03 in March, 1958-59 in March, 2023-24 in November. Several of these also see good snow in the main winter too. My guess is the MJO will set up to favor bigger storms nationally in something like an Oct 15-->Nov 30-->Jan 15-->Mar 1-->Apr 15 fashion for major events given the local Summer data supports unusual concentration of heavy snow events in Nov/Mar/Apr so far, more than the winter.

PDO has definitely weakened toward neutral in recent days. Net trend will be more positive. It's just whether that continues into actual positive territory and sustains. The North Atlantic on the other hand has also warmed up quite a bit.

July was similar nationally to 2002, 2003, 2006 more than the recent El Ninos, in terms of US temp profile. Recent events (2009-) were either much colder or had different spatial distributions for warm/cool areas. July is close enough to winter that you can probably infer meaningfully what will happen from the tele-connection set. This is July 2026: -EPO, -WPO, +PNA, -NAO, +AO. Of those, EPO was extremely negative. -EPO July, El Nino: 1957, 1963, 1969, 1976, 1979, 1986, 1991, 1997, 2002, 2003, 2006, 2018 -WPO July, El Nino: 1953, 1957, 1969, 1972, 1986, 1991, 2002, 2006, 2014, 2018, 2019 +PNA July, El Nino: 1953, 1958, 1969, 1979, 1982, 1991, 1997, 2002, 2006, 2009, 2014, 2015, 2019, 2023 -NAO July, El Nino: 1957, 1958, 1963, 1968, 1976, 1977, 1991, 2009, 2015, 2019, 2023 +AO July, El Nino: 1951, 1953, 1969, 1979, 1982, 1994, 2002, 2006, 2018 On the five signs, you only have a handful of El Ninos that are 3-4-5 match candidates. Those are bold above. 3 match group: 1953, 1957, 1979, 2018, 2019 4 match group: 1969, 1991, 2002, 2006 If you roll that nine year blend forward to winter - certainly somewhat more muted and different spatially then what I expect. The -EPO July is actually a fairly big warm signal for the following December in the Great Basin, weaker for the rest of the West to the NM/AZ border, and a weak cold signal for the SE US south of DC. That's like a +30% snowfall winter composite for me. All of those winters have fairly notorious off-season flukish snow events, which is a major feature of El Nino. You can get heavy snow in Oct-Nov and/or Mar-Apr in additional to decent snow in winter.

Closest simple match I could get for May-July US precipitation was a blend 1963, 1976, 1997, 2003 as a blend. A lot of years in winter end up coldest where Summer was wettest, 2009-10 was very cold in the East/South following a very wet May-Jul as an example. Most of the recent Ninos are much better than observed for the past May-July. The SE ridge is essentially displaced v. the blend, so the moisture in the SE has to be shifted to the NW, to include a bit more of the Plains and AZ, with the core dryness also centered somewhat NW of shown. For highs and rain locally, June 1-Aug 8 is almost identical to a blend of four El Ninos: 1953, 1958, 1994, 2023. Very hot, very dry. 2026 Jun 1-Aug 8 mean high: 94.9. Blend: 94.05F 2026 Jun 1-Aug 8 total rain: 0.80". Blend: 0.73" Interestingly, that blend is pretty consistently snowy in Nov-Jan locally, and then much more inconsistent Feb-Apr. But February is probably one of the warmest months on record for the US in that blend.

Last I looked, Jun-Jul precip totals here were tracking pretty close to 1982. That's generally a sign that pressure patterns and moisture levels are similar. Much colder/wetter than 2023 so far. 1 2023-07-29 T 0 2 1980-07-29 0.09 0 3 1958-07-29 0.36 0 - 1951-07-29 0.36 0 5 2011-07-29 0.39 0 6 1942-07-29 0.41 0 7 1995-07-29 0.52 0 8 2003-07-29 0.61 0 - 1947-07-29 0.61 0 10 1994-07-29 0.66 0 - 1993-07-29 0.66 0 12 1962-07-29 0.71 0 13 1977-07-29 0.73 0 14 2026-07-29 0.80 1 15 1982-07-29 0.86 0 16 2016-07-29 0.87 0 17 1953-07-29 0.88 0 18 1945-07-29 0.89 0 19 1971-07-29 0.93 0 20 1934-07-29 0.98 0 Highs are close too for July 1-29 1 2023-07-29 99.2 0 2 1980-07-29 99.0 0 3 2003-07-29 97.2 0 4 1978-07-29 97.1 0 5 1951-07-29 96.1 0 6 2016-07-29 95.9 0 - 1979-07-29 95.9 0 8 1994-07-29 95.7 0 9 1983-07-29 95.2 0 10 1963-07-29 95.0 0 11 2026-07-29 94.8 1 12 2011-07-29 94.7 0 13 1993-07-29 94.6 0 - 1982-07-29 94.6 0 - 1934-07-29 94.6 0 16 1954-07-29 94.4 0 17 2005-07-29 94.3 0 - 1981-07-29 94.3 0 19 1995-07-29 94.0 0 - 1953-07-29 94.0 0 21 1947-07-29 93.8 0

I can pull composites like these pretty quickly again now that I have my weighted analog system back, with he full package of predictors (ENSO, AMO, Solar, Modoki, PDO, ENSO prior yr, and monsoon rain for high/ridge positioning over the deserts). I took out 1997 because it was further along, but it only dulls the warmth a bit and you still have the general look.

Some of you are probably wondering how I ended up with my blend of 1963, 1982, 1997, 2004, 2015, 2018. If you look at Apr-Jun unweighted, its actually a very close match to 2026 for Apr-Jun, although running about 2F too cold in warm areas, in part because of how strong this event is, and how much warming has gone on since ~1995 the analog center point for time. You just have to make the north a bit colder, south warmer, and then the gradients are a bit more extreme.

The PNA is negative to barely positive in July in a lot of the big El Ninos on this data set. https://psl.noaa.gov/data/correlation/pna.data 2023: +0.65 2015: +0.29 1997: +0.10 1982: +0.49 1972: -1.13 July was negative in 2018, 1963, 1965, 2004, etc along with 1972 Was also negative in Aug 1997.

The Midwest Regional Climate Center snow maps for the years I picked. You can see generally the Northern Plains and Ozarks sort of get screwed for snow. Northwest does OK, especially if the PDO stays negative into Fall. Northeast does OK. Southeast & Southwest are generally furthest above means for snow. I can't get the maps to work for 2004-2005 and 1963-64. This is my attempt at drawing the consolidated snow maps. I put in a few example cities. So Richmond (~50%) is further above average than Philadelphia (~20%) and Philadelphia is further above average than Boston, which is a hair below (5-10%). Detroit and Chicago are around 20% below average. Topeka is right on the edge of the just below average/just above average zone (since it is low elevation, and blue western areas will be highly elevation dependent in a warm pattern for snow). Reno was something like 40% above average, Albuquerque around 20%. I'm using 1961-2020 as snow baseline. Billings is 3,000+ feet above sea level and around 15% below average. Those higher areas in the orange areas won't do too poorly, its the lower spots that get hosed. I may change my main analogs for temps, but I don't expect to change the snow maps. I'm pretty comfortable with this group for where moisture will be.

So the new AMO data is not de-trended like the old AMO data I used. If you de-trend it for time/global warming, correlation (r-squared) is over 0.8 between the two data sets for Nov-Apr. I also converted the PDO data back to the old Nate Mantua/JISAO methodology which ran about +0.5 more positive (2014-15 was +2.07 from Nov-Apr by that method, now shows as +1.59). Converted to the old methodology, where PDO was +2.07 and not +1.59 in 2014-15, I'd expect PDO in the -0.25 to +0.25 range for Nov-Apr (but it will show as -0.75 to -0.25 probably on the current methodology overall). AMO looks like it will be in the +0.20 range. Even with the de-trending, AMO was bonkers in 2023-24, +0.60 for Nov-Apr, warmest ever v. baseline and long-term trends. Part of why the 1972-73 pattern held (warmest NE 1/4 US, much colder elsewhere) but not the actual temperature pattern that year (much hotter). It's also really interesting seeing how weird the last solar cycle was in retrospect. We're quite similar to the down-slope of the cycle that peaked in 1968-1969 at 155 (we peaked at 149 in 2024-25). The QBO data stopped in Feb 2026. However, the six most similar years to 2025 for QBO, by objective absolute value measure are: 2014 then 1972 then 2021 then 1974 then 1979 then 1958. The winters starting 2015, 1973, 2022, 1975, 1980, 1959 are all +QBO winters in the +0 to +10 range. El Ninos in that QBO range are 1957, 1963, 1969, 1977, 1982, 1987, 1994, 2004, 2006, 2015, 2018. The Jamstec believe it or not has the Modoki index reaching +0.95 or so for Dec-Feb, v. +0.5 forecast for Dec-Feb forecast in July 2023 for Dec 2023-Feb 2024. I think it will trend lower, but its probably recognizing (correctly) that this event will decay rapidly in late winter from east to west, with the warmth maybe peaking in Nino 3.4 in Nov/Dec/Jan and then retreating to Nino 4 in Jan-Mar before collapsing in Summer. I'll guess +0.65 for the Modoki index. I lean toward something like this now that my old analog system is fully up to date. But I wouldn't be surprised to see a pretty cold stretch in the East later in the winter, probably late-Feb to mid-Mar for about three weeks if I had to guess when you do briefly see pretty ideal Modoki conditions. I think Dec-Feb overall is pretty warm almost everywhere, especially North, even with a few intense, but brief cold snaps. Winter ONI DJF AMO N-A Sun Jul-Jun ONI DJF(p) Modoki PDO N-A Monsoon QBO DJF Weight 1963 0.8 -0.036 29.1 -0.4 0.48 -0.88 5.06 4.9 x1 1982 2.3 -0.156 129.2 0.0 -0.49 0.95 3.84 10.9 x3 1997 2.3 0.220 54.9 -0.4 -0.66 1.24 6.58 -1 x1 2004 0.5 0.222 55.3 0.3 0.48 0.47 4.06 0.3 x1 2015 2.6 0.211 55.8 0.6 0.32 1.70 5.02 9.2 x2 2018 0.9 0.049 5.5 -0.8 0.41 0.51 4.84 8.8 x2 Blend 1.75 0.05 64.96 -0.09 0.03 0.81 4.69 7.29 x10 Target 3.20 0.15 75.00 -0.40 0.65 0.30 3.67 5 Long-term, since Nino 3.4 is a stronger warm signal in the North and weaker in the south, the composite should actually be more extreme than depicted by the blend I used - If you exaggerate the features of the winter to account for record setting warmth in Nino 3.4, I think you end up with something like this. Very warm in the North, still actually pretty snowy, and not really cold anywhere. I think we'll likely have a pretty cold December in the Southwest, and that's maybe the most extreme cold temperature month for the winter nationally. Sort of a joke map - but my current rough idea. Look at December - I think it will be warmer than this - by a lot in the warm areas and likely colder in the SW by quite a bit too. Now look at the Jamstec and the composite. The east-based look is real and present. But it will likely be breaking down by mid Nov to mid Jan. I'm not warm enough for the Nino zones - you can't get a +3.0 to +3.5C in Nino 3.4 with the anomalies. But I essentially was able to validate that the Jamstec isn't crazy. It has a canonical +PNA - very warm NW/cool SE. I don't quite think we get that with the AMO / PDO as they are.