raindancewx

I've basically finished my winter forecast using the updated analog weighting. Snow maps are a pain in the ass, I ended up looking at something like 40 or 50 towns to try to draw the boundaries correctly. I was hoping the ECMWF plume for Nino 3.4 would be out by now, but it isn't. If it isn't out tomorrow, will probably just post my forecast without it. Here is the subsurface trend for the past few months. The Nino 3.4 zone is completely filled up now with warmth. It does look to me like the warmth is moving East, so the Modoki look right now will probably fade later in winter, I think that is why my analogs are showing a big mid-winter thaw nationally for a month or so.

It looked like a slightly below average snowfall season for Denver when I ran my initial simulation. But that blend assumed a Neutral with warm waters in Nino 3.4 staying west of 140W. It seems like they'll spread east enough to be an El Nino. Ultimately I think its a fairly seasonal winter for temperatures for NM and CO, but with a handful of very strong, wet storms. If the MJO gets stuck in phase 8/1/2 in winter, like it is now, the current system set to be super cold for the time of year should show up again at some point, even if its not til March, but further South and more moisture. My original weighting was 1953-54, 1983-84, 1992-93, 1995-96 (x3), 2009-10 (x2), 2018-19, but I can't really justify keeping 1995 that strong, even though Summer 1995 and the QBO are similar to 2019. I'm planning to have my outlook out later this week - I have the new weighting with the same years, it just takes a while to manually re-do the maps and data mining.

It's late forming and not certain...but looks like an El Nino Modoki is forming. Fairly cold waters by Peru & the Philippines with warm water in Nino 4/3.4

There are no low-solar El Ninos after a low-solar El Nino for at least the last sixty years, some debatable cases in the 1950s - (1952 and 1953, but I consider 1952-53 to not be an El Nino). So going against it was pretty sensible. I still have this sense that the El Nino trying to form now will still fail somehow. That being said...my analogs last year were 1953, 1976, 1986, 1994, 2006 - 1976 and 1986 were both followed by El Ninos, and Summer has been fairly close to 1995 nationally. Last time a low solar El Nino followed a low solar El Nino was in the 1910s if you use 50-sunspots or less from July-June, in both El Ninos as the threshold. That would be the 1914-15 winter, when it hit -7F in January with 27.5 inches of snow in Albuquerque at an observation site lower in elevation and south of the current airport. I believe it is the snowiest winter here and across the Southwest generally since the 1890s. The 1913-14 winter actually has a passing resemblance to last year if you lag the timing so that's pretty interesting. The ECMWF plume update for October still isn't up from what I can see. Probably a big change in the forecast. The MJO similarities to 2018, 2012, 1994 even 2009, 2007, 2004 to some extent aren't real bullish for the NE for winter. 1995 had been a strong analog for me, but it is fading away now, the crappier snow winters for the NE are somewhat stronger analogs now, if an El Nino forms. My old analogs had the warmth reaching 140W in the tropical pacific in Nino 3.4, which is where it is now. It will probably continue to 100-120W over time - question is how soon. It's still very cold in the Eastern Tropical Pacific. The fade of warmth by Alaska is consistent with the PDO going negative by the way. That warm tongue from Japan to south of Alaska should not be there at all in a positive PDO year. Hard to get cold in the Southeast if that setup remains.

I'm about ready to wave the white flag on a Neutral at this point. My Neutral analogs included three El Ninos anyway, so I just have to re-do my forecast using the new weighting of the six years I like. Still want to see what the October ENSO plume from the ECMWF has. There are no warm-Neutral years in low solar since 1952-53, so I'd bump from 26.5C to 27.0C ish. The US temperature map doesn't even change that much, mostly December impacted, and then generally less blocking and less snow for the NE if I change the weighting. Nino 1.2/3 look like they're getting cold again this on Tropical Tidbits, that may make it to Nino 3.4 before the warmth below Nino 3.4/4 surfaces. Part of why I want to see the European - it did have the warm up in October in Nino 3.4 after all. Nino1+2 Nino3 Nino34 Nino4 Week SST SSTA SST SSTA SST SSTA SST SSTA 04SEP2019 19.8-0.6 24.9 0.0 26.6-0.2 29.1 0.5 11SEP2019 19.1-1.3 24.4-0.5 26.4-0.3 29.0 0.3 18SEP2019 19.1-1.3 24.2-0.6 26.5-0.2 29.3 0.6 25SEP2019 20.0-0.5 24.8-0.1 27.2 0.5 29.7 1.1 02OCT2019 20.0-0.6 25.1 0.3 27.2 0.5 29.7 1.0

I've been assuming Neutral, but we'll see what the weeklies do this week. If we spike to +1 or something then I've probably got to change my forecast to assume an El Nino. A low solar El Nino winter after a low-solar El Nino would be unprecedented for at least the last 90 years although 1953-54 is kind of close (9.5 sunspots, after a 26.9C ENSO). My hesitation is Nino 3 - there are no Octobers with Nino 3 as cold as it was last week on the weeklies (since 1990) that became El Ninos. It's possible some of the old Ninos did it prior to 1990. Nino 3 has a lot less warming long-term than Nino 4, so I look at it as a better indicator. The European also has a reversal/weakening SOI signal (high pressure by Tahiti lower by Australia) long term, which may weaken the pulses of warm water moving East soon. Sometimes the double El Ninos die in the Summer and then re-form in winter. 1976/1977, 1986/1987 had breaks. In 1992, the neutral-ish Nino 3.4 held on through winter before developing into an El Nino in early 1993 again. Given the Nino 3.4 match to 1992, I've been assuming the cool waters in Nino 1.2/3 would occasionally spill West and hold off enough warming for winter to be considered Neutral - even if Nino 3.4 is back up to 27.0C or by February.

Some of your old friends are showing up as MJO analogs, i.e. the MJO getting stuck in phase 1 in October (1994, 2012, and even 2018 are fairly similar):

It doesn't seem to apply outside the Boston to Philly corridor, east of the hills, but with last year included the "low solar El Ninos (generally)= fairly low snowfall" held up pretty well, and I had based that on the 1890s-2010s last year. This is total monthly sunspots July-June, divided by 12, for El Ninos in Boston. I've done statistical tests on this nationally, and solar stuff seems to impact precip/snow/temps when the annualized sunspot number centered on winter is 50 or less. El Nino Sun Jul-J Bos Snow 1899 18.2 25.0 1900 8.6 17.5 1902 18.7 42.0 1911 5.4 31.6 1913 7.4 39.0 1914 44.5 22.3 1923 14.6 29.8 1930 46.3 40.8 1953 9.5 23.6 1963 29.1 63.0 1965 37.1 44.1 1976 23.2 58.5 1986 19.1 42.5 1994 36.9 14.9 2006 20.1 17.1 2009 13.2 35.7 2018 5.5 27.4 Mean 21.0 33.8 If you do July-Jun for annualized sunspots, the highest snow in an El Nino with low solar in like 63 inches in 17 tries. So my point is if I'm wrong and an El Nino develops I'd actually expect my snowfall totals for the NE to be too high. I like 1992-93/1953-54 in some ways for precipitation more than temps, but you have some pretty warm spells in those periods if you how to time it right, Sept wasn't too far off nationally for temps from 1953 either. I'll show you guys my snow map next week - it is pretty detailed nationally. I don't see any reason to change it for now.

The matches for the three months show up as El Ninos because July was still warm in Nino 3.4, and so it looks like a developing El Nino. That's why by September alone, the matches don't look like an El Nino. Most of the modelling had been showing a warm-rebound in Fall since the Summer, I don't really the big deal. Nino 3.4 needs to reach 27.0C for an extended period just to be a weak El Nino, it has has been dropping in absolute terms and by anomalies since Summer. There is still cold water below the surface from about 120-140W, and Nino 1.2/3 are still around average, despite Nino 4 being very warm. It looks like a flat-Neutral to me by CPC designation, with some warm weeks and some cool weeks. Look at the Nino 4 years that are around 29.3C (29.1-29.5C) in September - its a very warm signal for the Northern US. (the image should say 29.1-29.5C in Sept) We went through this last year, and Nino 4 is nearly identical. Nino 4 is a only weak signal for warmth in the North, but near record values with a weak signal still implies warmth. Also, the big -SOI in September supports a warm East in December. I personally don't care if an El Nino develops, Modoki El Ninos and +PDO Neutrals behave similarly here. With last year included, the NE corridor still tends to not do well for snow in low-solar El Ninos if that is what is to happen. These are the closest Nino 4 years for September (29.1C-29.5C). They are all El Ninos - but keep in mind Nino 4 is warming much more rapidly than the other Nino zones. That's part of why you guys in the East have had such trouble getting a cold December since Nino 4 is correlated to December in the East for temps. 1989-2018 Septembers in Nino 4 are 28.75C on average, compared to 28.11C in 1950-1979. If the 2017 Nino 4 reading was applied to the older average, it'd be +0.6C, but a La Nina still developed, one that had the coldest Nino 1.2 readings for a while: 30 to 40 years in several months. My interpretation is if we had data for 2005-2035 in Nino 4, you'd find it was only +0.2C, instead of +0.55C. There are plenty of years when Nino 4 is +0.2C against the "centered climate average", that do not become El Ninos. 2004 29.45 1997 29.44 2006 29.40 2002 29.37 2018 29.34 2009 29.33 2019 29.32 1994 29.30 1969 29.16 2014 29.16

I went with a warm AMO, warm PDO, Neutral/"cooler El Nino Modoki" look for my forecast, with low-solar, and ENSO order factored in. It's actually a pretty snowy winter for a lot of the US if I'm right. Well...first flake to last flake is snowy. Nino 3.4 is remarkably close to 1992. It's actually pretty difficult to find a cooling Nino 3.4 that is still above average for JAS historically. https://origin.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/detrend.nino34.ascii.txt 2019 6 28.19 27.65 0.54 2019 7 27.65 27.26 0.39 2019 8 26.91 26.91 0.00 2019 9 26.69 26.80 -0.11 1992 6 28.30 27.60 0.70 1992 7 27.51 27.19 0.33 1992 8 26.91 26.84 0.07 1992 9 26.66 26.78 -0.13 My analogs have a dry slot from east-TX to Michigan for the winter, but the general idea here seems about right to me. https://www.onthesnow.com/news/a/635392/winter-19-20-long-range-weather-forecast For me, 1992/1953/1953 as a blend has been fairly close for rainfall patterns, as has 2004/1989, 1953/2015. 1992 is my wettest winter on record by a country mile (3.92"!) but 1953 is very dry. My interpretation is it is a predominantly dry pattern with extremely wet storms sprinkled in occasionally. Something like seven real storms for me between Nov-Apr, but they'll all be pretty big.

My analogs had Dallas slightly above average for snow I believe. Something like +10% or +20% if I'm remembering right. I'll put out my winter forecast here in a week. It's basically done, just want to see a couple more things. The MJO is moving erratically - I want to see if that actually persists or not.

I have a pretty snowy winter for the US generally this year. Our mountains here won't have snow on them from mid-October to mid-June like last year, but the mountains never really stopped dropping into the teens/20s this Summer here, there were frosts in populated towns in July on some nights. The entrenched night time dryness favors high snow ratios out here, even by mountain standards, even though the precipitation signal (for totals) isn't great. Ski visits and resort level snow were the best since 1997-98 out here, with 140-180% of normal values at that level, including early openings and late season extensions (Nov 16-Apr 10 or so, v. the more typical Nov 23-Apr 3). I have a sharp gradient between Northern Maine and Boston for anomalies this year, essentially the reverse of last year, with Boston seeing more relatively, and Maine seeing less relatively as you head North. US should be much drier than last year, but I think the storms that come through will often be extremely powerful. Something like +30% by Boston v. -30% by Caribou if I remember correctly, for snowfall.

Year DJF JFM FMA MAM AMJ MJJ JJA JAS ASO SON OND NDJ 2010 1.5 1.3 0.9 0.4 -0.1 -0.6 -1.0 -1.4 -1.6 -1.7 -1.7 -1.6 2011 -1.4 -1.1 -0.8 -0.6 -0.5 -0.4 -0.5 -0.7 -0.9 -1.1 -1.1 -1.0 2012 -0.8 -0.6 -0.5 -0.4 -0.2 0.1 0.3 0.3 0.3 0.2 0.0 -0.2 2013 -0.4 -0.3 -0.2 -0.2 -0.3 -0.3 -0.4 -0.4 -0.3 -0.2 -0.2 -0.3 2014 -0.4 -0.4 -0.2 0.1 0.3 0.2 0.1 0.0 0.2 0.4 0.6 0.7 2015 0.6 0.6 0.6 0.8 1.0 1.2 1.5 1.8 2.1 2.4 2.5 2.6 2016 2.5 2.2 1.7 1.0 0.5 0.0 -0.3 -0.6 -0.7 -0.7 -0.7 -0.6 2017 -0.3 -0.1 0.1 0.3 0.4 0.4 0.2 -0.1 -0.4 -0.7 -0.9 -1.0 2018 -0.9 -0.8 -0.6 -0.4 -0.1 0.1 0.1 0.2 0.4 0.7 0.9 0.8 2019 0.8 0.8 0.8 0.8 0.6 0.5 0.3 0.1 Nino 3.4 is still tracking very close to 1992 for recent months. https://origin.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/detrend.nino34.ascii.txt 1992 6 28.30 27.60 0.70 1992 7 27.51 27.19 0.33 1992 8 26.91 26.84 0.07 1992 9 26.66 26.78 -0.13 2019 6 28.19 27.65 0.54 2019 7 27.65 27.26 0.39 2019 8 26.91 26.91 0.00 2019 9 26.69 26.80 -0.11 For July-September, this is what I have as the top six objective matches in Nino 4, Nino 3.4, Nino 3, Nino 1.2: 1994, 2001, 2018, 1990, 2017, 1977. For September alone: 2005, 2013, 2001, 1990, 1981, 1994 Keep in mind, Nino 4 was 29.3C in September - as warm as last year. Very hard to get a cold December in the East when Nino 4 is very warm, even if the NAO is negative. Against 1951-2010 averages, this is what you get for the four zones in September: Nino 4: +0.92C Nino 3.4: +0.14C Nino 3: -0.16C Nino 1.2: -0.42C Nino 1.2 is as cold as it was in 2017 in September, but Nino 4 is a warm as it was in 2018.

European has the drop in Nino 3.4 for September.

A lot of SE New Mexico locales had near record rain the past few days. I believe Roswell had about four inches in a single day.

The subsurface data for 100-180W, 300m down, has finally broken (warmer) away from 2017. This is the closest match I could easily come up with for July-Sept. https://www.cpc.ncep.noaa.gov/products/analysis_monitoring/ocean/index/heat_content_index.txt 100-180W July Aug Sept 1984 -0.25 -0.22 -0.37 2003 0.53 0.03 0.10 2013 0.41 0.32 0.38 1984 -0.25 -0.22 -0.37 2003 0.53 0.03 0.10 1984 -0.25 -0.22 -0.37 Mean 0.12 -0.05 -0.09 2019 0.13 -0.08 -0.05 1984 (x3), 2003 (x2), 2013 as a blend actually kind of looks like what I have for winter in the US, but from completely different analogs. Look at how much colder 2017 was by this point in the subsurface data. If you plot 1979-2018 September subsurface data against Nino 3.4 DJF temps, you'd expect about 26.6C give or take a few tenths of a degree at pretty high certainty. 100-180W July Aug Sept 1984 -0.25 -0.22 -0.37 2003 0.53 0.03 0.10 2013 0.41 0.32 0.38 1984 -0.25 -0.22 -0.37 2003 0.53 0.03 0.10 1984 -0.25 -0.22 -0.37 Mean 0.12 -0.05 -0.09 2019 0.13 -0.08 -0.05 2017 0.16 -0.40 -0.79

These are the top SOI matches for July/Aug/Sept since 1931: 1946 1948 1951 1963 1969 1991 Outcome as a blend is a cold NW / warm East look for October, which doesn't seem too far off from reality, the models / CPC and my analogs all had that too. 1963-64 is a pretty extreme winter in December for the US, but it should drop out as a top SOI match soon, unless the SOI is around -15 again in October. If you look at October 1963, it is also the mother of all blow-torches for everywhere but California, so not exactly likely. Like last year, different MJO background in all likelihood.

The Canadian update for 9/30 still looks dramatically over cooled after October for the global oceans, with nearly all of Earth's cold/in blue. However, the air temperature maps and precipitation maps still look reasonable to me. October is depicted to retain the coolish Nino 1.2/3 with half of Nino 3.4 cold and half of it warm, and then Nino 4 warm. The NAO went pretty negative in late September, will be curious to see what the final NAO value is for September. The PDO also looks less positive for September. If Nino 1.2 is near average in October but the PDO is still positive in September, it is likely that the PDO will remain slightly positive through winter, but not probably positive enough to drive cold into the far SE US persistently. I have my snow map ready for the winter forecast - a lot of places in the West that had heavy snow in late Sept/early Oct show up in the analogs. If I'm right, most of you will be pretty happy, outside of some areas in the Ozarks and Lakes that may get jacked. Neutral composites for snow are pretty interesting, but do "bend" a bit to ENSO order if that makes any sense. Generally speaking, I think the heavy snows move around a lot this cold season. Heavy snow events in Fall/Spring for the NW and Northern Plains, with SW in Nov-Dec and Feb-Mar, and then the NE/MW probably due well in Jan/Mar. Anyway, I'm fairly excited, it looks like a fun winter nationally. I do think parts of the West will be pretty warm, but not to the extent of 2013-14, 2014-15 or 2017-18. I looked at snowfall in my analog years for around 25 cities in the lower 48, most places do well or are near average, not many that are well below-average.

ONI weeklies all shot up again, and the weekly ENSO update PDF from CPC/NOAA has the subsurface above average again for 100-180W down to 300m. Nino1+2 Nino3 Nino34 Nino4 Week SST SSTA SST SSTA SST SSTA SST SSTA 04SEP2019 19.8-0.6 24.9 0.0 26.6-0.2 29.1 0.5 11SEP2019 19.1-1.3 24.4-0.5 26.4-0.3 29.0 0.3 18SEP2019 19.1-1.3 24.2-0.6 26.5-0.2 29.3 0.6 25SEP2019 20.0-0.5 24.8-0.1 27.2 0.5 29.7 1.1 The weeklies imply 26.68C or so in Nino 3.4, which is +0.15C for September, against 1951-2010 averages. Nino 1.2 would be -1.0C, Nino 3 -0.1C, Nino 3.4 +0.15C, and Nino 4 +1.0C, all against 1951-2010, assuming the monthlies match the weeklies - and they tend to except in Nino 1.2. The SOI finished around -13 in September. CPC uses different averages for differet time frames, my table below does not. The table linked here goes to the 1800s for estimating ONI from different data. https://www.webberweather.com/ensemble-oceanic-nino-index.html Using actual SST data, not anomalies from CPC, 1992 remains a very strong match to 2019 so far in Nino 3.4 My winter analogs also look like what CPC is showing nationally for October, which I find encouraging.

The 2017-18 La Nina was developed by this point. Right now, 2019-20 still looks Neutral. Those waters north and northwest of Australia should really inhibit any major cool down in the Nino zones. Tropical Tidbits has Nino 3.4 back to near 0, with Nino 1.2/3 warming back to cold-neutral from La Nina territory. There is a lot of warmth set to come up in the West now too if you look at the 9/25 frame.

Honestly, you guys seem to be over-reacting a bit to last year. Conceptually, if you split the lower 48 into four regions, the strongest indicators for a cold winter are something like: NE 1/4: -NAO, or a Modoki El Nino with high solar (2014, 1977, 2002, 1968, etc) SE 1/4: +PDO. Much stronger correlation to southern cold than any ENSO state if you look historically. 1983-84 and 1995-96 were both +PDO years for Nov-Apr. SW 1/4: An El Nino (>=27.0C) winter after a La Nina winter (<=26.0C). This is actually becoming rarer as a lot of years are cold Neutrals now instead of weak La Ninas. (2018, 2009, 2006, 1976, 1972, 1951, 1965, etc). In the SW, warming up from a cold Nino 3.4 dramatically y/y is a pretty strong indicator for cold if in the initial year is cold. Even 2011, a La Nina but much warmer than 2010, but fairly cold in New Mexico is an example of this, despite not being an El Nino. NW 1/4: A La Nina after an El Nino (2016, 2010, 2007, 2005, 1998, 1995 etc). I believe each of these since 1995 has been cold somewhere between the Dakotas and N. California, even 2005-06 was relatively cold in this zone v. the rest of the US. Last year, conceptually fit the SW cold pattern. However, if you recognize that the SOI jumped to +9 in Dec 2018 after being -3 in Dec 2017....that's actually close to a "La Nina after an El Nino" jolt to the SOI. Shockingly, the atmosphere responded and it got stupid cold in the NW & Plains in February, it snowed in Malibu, Billings had its coldest month since 1936, and so on. For these splits, the r-squared values approach 0.5 over long periods for indicating cold. It isn't that there aren't exceptions, but I tend to view these relationships as robust, as you get the opposite effects in each region when the relevant sign is reversed. ENSO has always been over-rated as a live indicator for seasonal stuff, it's more important for telling you where things are coming and where they came from.

The IOD (Indian Ocean Dipole) is pretty interesting for how the Tropical Pacific looks right now. The set up in the Indian Ocean is more consistent with an El Nino. Part of me thinks this event will transition back to an El Nino, later in 2020 and then after that we'll get 2-3 La Ninas in a row, which will be terrible for most of the West. The SOI is still around -15 for September, which is broadly speaking a wet signal for the West but not so much where I am. The NAO looks like it will crash again for a bit late this month. I'm curious to see what the final NAO value is for September. My hunch is the NAO will be more volatile than predominantly negative or positive this winter. My tentative analogs have one -NAO month, one near average, and one +NAO month. We will see. September relative to March and May relative to April as a blend works pretty well in figuring out the predominant NAO state the next winter I find.

I was actually looking at snow for the upcoming cold season in New Mexico, and the higher elevation areas come out pretty well, the signal goes from below average snow at valley level to above average at the mountain tops. Highest terrain may be up to 20% above average. The signal I get for Albuquerque is for a lot of cold nights despite somewhat warmer than average highs. Last year, the NW 1/3 and Gila (SW NM by AZ) did well, generally normal to +80%, with the southern valleys and SE part of the state well below average. This is probably what I'll use for winter forecast, pending any last minute surprises in September (say, 2.4 inches of rain from now to 9/30 in Albuquerque).

https://www.cpc.ncep.noaa.gov/data/indices/wksst8110.for Nino1+2 Nino3 Nino34 Nino4 Week SST SSTA SST SSTA SST SSTA SST SSTA 28AUG2019 20.2-0.4 24.6-0.3 26.5-0.2 29.2 0.6 04SEP2019 19.8-0.6 24.9 0.0 26.6-0.2 29.1 0.5 11SEP2019 19.1-1.3 24.4-0.5 26.4-0.3 29.0 0.3 18SEP2019 19.1-1.3 24.2-0.6 26.5-0.2 29.3 0.6 Nice rebound in Nino 4. Nino 3.4 hanging tough. 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 A warm September in Nino 4 is a strong cold signal for a lot of the US, and that's broadly what my analogs have for October.

One of the old Jamstec runs had a cold Nino 3 / 1.2 peaking in SON, before decaying in winter. That seems more or less right to me. I don't disagree at all about Nino 1.2/3 remaining cold. But the +2 and -2 cold pools on the image above are both at the edges of Nino 3.4, at similar depth and similar magnitude. So I don't see any reason for warm west / cold east to change. The trick is Nino 3.4. This is the July run for SON. Looks like right now in the Nino zones. But then it had a decay in DJF. Nino 1.2 has a tendency to predict PDO changes, so the PDO zone is less positive than depicted, but otherwise, this was a good map. I think the weeklies will show some warmth recovery in Nino 4 and maybe Nino 3.4 this week, probably not to the extent Tropical Tidbits has though.