raindancewx

Ignore this - image is from last year. I don't pay for Weatherbell. But looks like they changed their winter outlook again. I'm colder in the West, similar in the middle of the US (I think Texas could legitimately be +5 in a lot of spots) and then warmer in the East.

In the November 1950 storm, you had heavy snow on a south wind in Pittsburgh and a heavy rain event in Buffalo on a north wind, you can get all sorts of crazy stuff if you have a strong high and low interacting in unusual ways. This is from the Washington Post on the 1950 storm For instance, in Pittsburgh, heavy snow was falling in 21-degree air … then as winds shifted from northwesterly to southeasterly … the temperature plunged to 9 degrees, with the arrival of the cold front from the southeast! And while Pittsburgh was at 9 degrees (and heavy snow), hundreds of miles to the north, Buffalo was experiencing a balmy 54 degrees, heavy rain and hurricane-force gusts. It’s hard to believe that the District’s temperature was nearly 50 degrees warmer than Atlanta’s!

I think you'll see the magnitude of warmth on the CFS shift a lot cooler in the next few days for December. Spatial arrangement may be about right.

00Z run 11/26 from the GFS has 40 hours of snow for New Mexico starting around six days from now. It's basically dead on to the SOI crash time frame. I'm sure it will change, but the trend is good.

To me La Ninas are Atlantic driven. If we're really moving to the cold AMO, and I still think we are going to see it mid-decade - then we should start to see much better (colder) La Ninas nationally. 1959 if it verifies was about 4-5 years before the flip to the cold AMO, and that's about where we are now, given the flip to warm was 1995. The hurricane seasons from 1954-1964 are kind of similar to now too, with Hazel, Donna, Hilda, etc. 1961 had a huge system hit Central America super late like this year, and has the QBO/ENSO match you'd want if you like analogs off that stuff. I'd imagine you had (cyclical) low sea ice extent in those Falls too, although still more than now.

I name the winters by the year they start. My actual winter blend was 1995, 2003, 2007, 2012, 2019. I'm less confident on the weighting I used (2007 x5, 2012 x2, others x1), but unweighted it worked well in October. For 11/1-11/25, the actual high is identical to 2007 in Albuquerque (63.4F), and when I rank all years in the past 100 for precip and highs for June-Nov, 2007 is easily the best match overall out here. This is what I had. https://www.scribd.com/document/479516526/Winter-2020-21-Outlook There are already some issues with it: I assumed an ACE of 135, and based some of the pattern idea off 95-175 years. But the ACE is more like 180 now. The raw blend also had 4.8 inches of snow where I am, but we had 4.4 inches in October (all-time monthly record), and statistically we're far more likely to get multiple months topping four inches of snow here in one Oct-May season if we top 4 inches in an "off-season" month, i.e. Nov, or Mar-Apr, it's a common El Nino trait here. A lot of volcanic and Neutral winters here see huge snow dumps in November, so I kind of wonder if the October snow is something similar with aerosol stuff from the western fires this year - but I doubt it is that. The October system had incredible moisture too, a lot of the west side of the city had 8-14 inches of snow after being near 80 degrees a few days before. Since the long-term snow average is 9.6" but we average 6.2" in an average Nov-May La Nina, it's pretty likely we'll finish with above average snow here (~55/45 using La Nina years). Of course, there is no La Nina since at least 1930 with 4 inches of snow here in an off-season month, so we could also destroy the average pretty easily. A lot of La Ninas (~1/3) finish under 4.4 for the entire cold season, including 2017-18, 2016-17, 2010-11, 2008-09, 2005-06 just in the last 20 years. To me, the low-sea ice similarity is uncanny to November so far - Same for the +NAO La Nina patterns in November - that's not real typical. It's 1956, 1999, 2007, 2011 as a blend if you use only top third +NAO monthly values with a La Nina. That goes to a somewhat similar look to how December looks.

My idea for the winter has always been that you need a couple years with a warm North Pacific in the analog blend to match how the oceans look. I added in 2012 and 2019 to do that, since they both have low sea ice, and were colder by Peru than the date line at the equator in the Fall. In 2019 specifically, you actually have the warm patch at 45N/150W just like this year does. It's a stormy volatile pattern that ends up somewhat warm in the East and somewhat cold in the West in December. That's been the idea. Everything I can think of supports it: November +NAO La Nina supports it, low-sea ice extent cold ENSO supports it. Both of those actually look like November quite strongly. The "desert ENSO constant" supports it. Essentially if you do the Koppen classification math out here for calculating aridity, you'll evaporate/lose 3x the amount of moisture that falls in a July-June La Nina, so if you have exceptionally hot/dry conditions in Summer, you need a lot of help to move wetter toward that constant, and it will tend to come in winter, since La Nina is strongest as a dry-signal here in Spring. The 1959-60 winter has some fluky snowstorms pretty far south (including two feet where I live in December). My guess would be Dec and March would be the time frame for good eastern snows if 1959-60 ends up a good match to the winter. February 1960 is extremely cold nationally, and March 1960 is extremely cold in the East, so I don't really expect it to be a good match overall. It looks pretty reasonable for December though. It's pretty similar to December 1988. On the CPC maps, I think 1988 is probably a better match anyway. When I did my forecast I gave a long list of how I tiered analogs. The B tier analogs like 1988 should be good temperature matches for a lot of the US a lot of times, I'm just less sold on them for precipitation matches. I've been pretty happy with either 1995 or 2007 as the correct match for Fall for a while now, even if the timing is usually off. November 2007 definitely looks close to this year.

We'll see. Even from 48 hours ago, there was nothing like this depicted on the models: When I did my outlook, it seemed like 12/10 would have been the time frame for a big December system, but we seem to be getting a lot of things that happened similarly in 2007 at time frames that are 2-3 weeks off. There were big drops in the SOI late Nov and Dec 2007, but not all of them preceded big systems. It's definitely a decent indicator for low passages but not as reliable for actually identifying storms with...moisture. In 2011, it actually got so cold down here that a couple systems missed northern New Mexico because they were too far south, some of the things I look at are somewhat similar to 2011, have to watch for that too.

December never looked particularly terrible to be me for the East. Still think it's warm overall, but it's warm + snowy more than warm and boring. I've been looking hard at 1959 lately - that's a near La Nina with a very warm NE Pacific. There is definitely resemblance to early December 1959 in the current depictions by CPC. I'd say 12/1-12/10 overall looks a lot like 1988 too as currently depicted - both of those years still finish warm in the East but they're much better patterns for snow than a normal warm month would see. I think there are some SE snowstorms or ice storms this month down to I-40, but we'll see.

The first ten days of December as depicted now looks like 12/1-12/10 in 1988. Still a warm month in the East / cooler West by the end.

I wanted to give an update on this idea because I consider it theoretical/experimental with so few years. It appears to be working. The low sea ice extent cold-ENSO years are much closer to November 2020 (another low-sea ice cold ENSO), and also much closer to the CFS idea for December 2020 and the +NAO November La Nina look for December than the higher sea ice years. The higher sea ice years are too cold in the East. The big area of very warm November conditions has appeared in the right spot (AZ to Michigan) at the right magnitude, with a cool spot by the NW. Not perfect, but close, and definitely closer than the higher sea ice years.

We've actually been getting pretty decent snowstorms this Fall, hopefully the big warm ups that have melted most of it are done for a while. My winter outlook had 20 inches of snowpack at Taos Powderhorn entering 12/1, we may actually beat that with the coming storms.

One year I've been looking at a lot lately in 1959. The NW Pacific is cold that year, but you have a very warm NE Pacific with a pretty cold ENSO event. If you blend in years to warm up the NW Pacific with 1959 it's not that hard to get a map resembling the oceans today but less extreme. More importantly though, early December 1959 is dead on to the CPC forecast. That map has issues for the La Nina intensity, and the far northern oceans are different. But the MJO / tropical zones are are in the correct spots, the little SE to NW warm zone from India to Australia west of a cool patch south of Indonesia is there, the cold ENSO and cold tropical Atlantic are there. The warm patch into the NE Pacific is too far SW, but the real issue is that cold spot in the NE Atlantic. If you try factor in the northern oceans, this probably a bit better with 2011 in there to fix the north Atlantic in particular (it has the same cold blotch by Greenland as 2020), while 2008 fixes the NW Pacific.

NAO in November looks like it finishes around +1, essentially splitting the difference between 2007 and 2011, in terms of recent +NAO Novembers. The designation used here is top third for 1950-2019 monthly readings in November. Positive NAO November La Ninas are pretty rare - but they look a lot like what the CFS is forecasting currently for December, but with less insane warmth everywhere. The right image is also close to the "lowest sea ice extent" cold ENSO composite of 2007, 2011, 2012, 2016. One thing I wanted to mention - and it could be a coincidence - the +NAO November La Nina all evolve to much warmer ENSO events the next winter including 1949-50 where I didn't have NAO data but matched the pattern. The neutral November NAO La Nina winters vary a lot with years like 1954 and 2017 both included, but overall look closer to a cooler version of the pattern on the right. March still fits the pattern of November being a good indicator too -

How did you all do in December 1959? The CPC outlook for week one of December is basically a spatial repeat of that week. Very snowy month here.

On the old MEI data, 1954, 1964, 1988 are probably the closest matches to 2020, along with 2007. Snow patterns that are similar nationally in the Fall have tended to "look" like 2007 (it's like a blend of 1947-48/1996-97) somewhat after Nov 20, with the Plains usually getting hosed for snow somewhere. It's a mixed signal for most of the US, but I'd say overall still favors New England, Midwest & West, with NYC-Chicago on the edge, and areas south below average.

This is what I have if you're curious for a SW perspective on national trends. It's pretty amateurish compared to yours but I do have the advantage of knowing what tends to precede the better winters here. https://t.co/HJU1iBGTa4?amp=1 What I had for New Mexico is pretty similar to what you just said - relatively normal winter for precipitation, and then a very hot/dry Spring, probably starting late winter. The best indicators for La Nina here seem to be the ACE index for mid-winter highs (12/16-1/15 is very well correlated here to ACE) and actually September rainfall is correlated somewhat well to Oct-May precipitation believe it or not. September 2017 had 2.20 inches, and super high ACE and then there was a literal 96 day period without precipitation starting in October, with perpetual near record warmth. The current Fall is mostly warm but with strong recurring cold shots. Albuquerque only finished with 0.62" this September, which is low for a La Nina, and favors pretty normal total Oct-May precipitation (we get about 4.0" from Oct-May in a normal year). I've also found that since 2007, the cold-ENSO years with <=4.3 million square kilometers sea ice extent in September have all been cold or at least near normal in the Rockies or West, somewhere, while the years higher than that have been warm in the West, and cold in the Plains or Eastern 1/3 of the US.

The subsurface reading for November 2020 should be around -1.1. If you blend in the subsurface reading for Sept-Nov, it looks just like November 2020. The right image is how December would look. I'm assuming a lot of the reds down here burn off with the cooler highs this week, and the map of 2020 moves toward the subsurface blend.

Back to the prior level after the hiccup last week. Nino 4 didn't change at all though. Nino 1.2 is now closer in 2010 than in 2007. Nino 4 is still pretty warm for cold the middle zones are currently. Nino 1.2 warmed up in 2007 rapidly in December and should come back as a better match than 2010. Right now, 2011 is closer than 2007 or 2010. Nino1+2 Nino3 Nino34 Nino4 Week SST SSTA SST SSTA SST SSTA SST SSTA 07OCT2020 19.5-1.2 23.4-1.5 25.5-1.2 27.8-0.8 14OCT2020 19.6-1.2 23.6-1.3 25.3-1.4 27.8-0.9 21OCT2020 19.9-1.1 23.8-1.2 25.3-1.4 27.8-0.8 28OCT2020 20.0-1.1 23.5-1.5 24.9-1.7 27.9-0.7 04NOV2020 20.3-1.0 23.6-1.3 25.1-1.5 27.8-0.8 11NOV2020 21.2-0.3 24.1-0.9 25.7-1.0 27.9-0.7 18NOV2020 21.0-0.7 23.6-1.4 25.2-1.5 27.9-0.7 10OCT2007 18.8-1.9 23.3-1.6 24.9-1.8 27.7-1.0 17OCT2007 18.6-2.2 23.5-1.4 25.3-1.4 27.9-0.7 24OCT2007 19.2-1.8 23.5-1.4 25.3-1.4 27.8-0.8 31OCT2007 19.7-1.5 23.2-1.8 25.2-1.5 27.8-0.8 07NOV2007 19.3-2.1 23.0-2.0 25.0-1.6 27.5-1.1 14NOV2007 19.2-2.3 23.2-1.7 25.2-1.5 27.3-1.3 21NOV2007 19.8-2.0 23.2-1.8 24.9-1.7 27.1-1.5 05OCT2011 19.8-0.8 24.1-0.8 25.7-1.0 27.7-0.9 12OCT2011 19.7-1.0 24.1-0.8 25.8-0.9 27.9-0.8 19OCT2011 20.0-0.9 23.9-1.1 25.8-0.8 28.1-0.6 26OCT2011 21.2 0.2 23.8-1.1 25.4-1.2 28.0-0.7 02NOV2011 20.5-0.8 23.9-1.1 25.6-1.0 27.9-0.8 09NOV2011 20.5-0.9 23.9-1.1 25.7-0.9 27.9-0.7 16NOV2011 20.5-1.1 23.8-1.1 25.6-1.0 27.9-0.7 06OCT2010 18.7-1.9 23.2-1.7 24.8-1.9 27.1-1.6 13OCT2010 18.9-1.9 23.0-2.0 25.1-1.6 27.1-1.6 20OCT2010 19.1-1.8 23.2-1.7 25.1-1.6 27.0-1.7 27OCT2010 19.8-1.3 23.6-1.4 25.2-1.5 27.0-1.6 03NOV2010 19.7-1.6 23.4-1.6 25.2-1.4 27.0-1.6 10NOV2010 19.5-1.9 23.5-1.5 25.2-1.4 27.2-1.5 17NOV2010 20.2-1.5 23.5-1.5 25.0-1.6 27.1-1.5 24NOV2010 20.6-1.3 23.3-1.7 24.9-1.7 26.9-1.7 To me it looks like the subsurface is continuing to warm too, as the warmth below the western zones has had some success moving east recently.

"La Ninas are warm and dry in the Southwest". Well... My analogs had 0.40" total precipitation in November, with a bit of snow. So it'd be nice if this verified. The timing/precip amounts look fairly similar to one of the late November storms in 2019. That doesn't really surprise me - in the records here heavy snow in November is highly clustered in short periods, and then it disappears for decades. So would not be shocked if we see a big snow storm. Still like the 12/1-12/3 period for a big storm too given the huge SOI crashes of late.

During the 2007-08 La Nina, you had several huge SOI crashes that preceded some big storms in the West. The crash recently is reminiscent of this period in 2007 particularly - although somewhat earlier. It's worth noting for those who like 2010, that La Nina had almost no -SOI readings at all from Sept-Dec, while this year has seen them semi-regularly, as in 2007. Think you have to assume the subtropical jet won't be super weak despite the La Nina strength. Date Tahiti (hPa) Darwin (hPa) Daily Contribution 30 day Av. SOI 90 day Av. SOI 23 Nov 2020 1010.25 1008.60 -7.83 6.36 7.89 22 Nov 2020 1010.60 1008.90 -7.51 6.66 7.94 21 Nov 2020 1011.63 1008.30 2.86 7.21 8.15 20 Nov 2020 1013.38 1007.85 16.86 7.44 8.23 19 Nov 2020 1014.15 1008.10 20.17 6.99 8.14 2007 325 1011.01 1007.15 6.23 2007 326 1010.69 1006.85 6.11 2007 327 1012.06 1005.65 22.46 2007 328 1014.19 1007.00 27.42 2007 329 1013.41 1009.25 8.14 2007 330 1012.05 1009.80 -4.01 2007 331 1011.34 1007.85 3.88 2007 332 1012.91 1006.95 19.59 2007 333 1013.66 1009.50 8.14 2007 334 1013.05 1010.90 -4.65 The 332 to 334 represents 33 to 31 days before the end of the year, i.e. 11/29 to 12/1. The rule for the SW is a storm will appear ten days after - so 12/10-12/11 2007 saw 150% of normal (monthly) precipitation for Albuquerque. Will be very curious to see if we see a major storm in the 12/1-12/3 period this year. The big drops that go positive to negative are usually the strongest storms out here.

The years that are similar to snow this Fall since 1930 have generally been good for the West after the Fall too. A blend of 1947/1996 looks a lot like 2007, which I think will be a decent analog for snow the rest of the way (it hasn't been so far). The three days ending November 23 have each seen big SOI crashes (10+ points) in a two period - have to watch 12/1-12/3 for a big system or systems.

Eyeballing past years, I'd say the top snowfall matches to Sept 1-Nov 20 2020 are 1946, 1947, and 1985, with 1996 and 2011 the best matches of the past 30 years. Those are some pretty good snow years out here, except for 1946.

Are you the Jim Sullivan here? My main issue with 1988 has always been the high solar activity and to a lesser extent the super active monsoon in the SW US. For 1999, you had a very cold Summer in the Southwest and also very high solar activity. July-June years in the desert southwest when filtered by ENSO tend to see ~constant rates of incoming precipitation/outgoing precipitation (evaporation and sublimation), so in the years with well below-average evaporation in the Summer in La Ninas, we tend to suffer for it with hot or very dry winters, which is what happened in 1999, but also 2005, and 2017 to some extent. The hottest driest Summers in La Nina tend to have the opposite effect, with more moisture and/or cold necessary to offset the extra evaporation in Summer. Those years include 2007, 2011, 2016 - all very hot and/or dry Summers that were either cold and wet (2007), mild and wet (2011), or warm but very wet (2016). These patterns extend back to older years I find - with 1933 fitting the first pattern, and 1954 fitting the second as examples. Tends to work with cold-Neutral years too (2012 fits the second pattern, 2013 the first). I guess my point is I like that you have some wetter years in there for the SW winter, but I think it may be somewhat colder than people expect too. The +NAO La Nina November pattern (like 2011 say) is generally cold at least in December in the SW.

My issue with the 1970s La Ninas, almost all of which were cold here and/or snowy here, is you have a much colder North Atlantic and North Pacific, and generally low ACE values. 1970 and 1975 are both very low ACE years particularly for a La Nina. Atlantic ACE is loosely correlated with the AMO so that makes sense. Outside the NW US, it really does seem like the Atlantic drives patterns in La Nina winters, unlike in El Ninos where the Pacific matters far more. I'm not a huge fan of 1999 either. I'm actually pretty curious to see how November finishes because it's pretty tough right now to get a good spatial match for it. The six years I mentioned as good matches through 11/15 are really not all good matches anymore. 1978 turned very cold by this point in the Northern & Western US as an example. Here is the same analysis but six days later. Here is a look at the unweighted six year blended highs - 1964, 1975, 1978, 1987, 2011, 2015 for ten US cities 11/1-11/21: Boston: 58.0F v. (54.0+63.9+54.5+53.8+59.0+58.6)/6 --> 57.3F Richmond: 67.2F v. (69.3+72.6+67.0+64.6+64.1+65.8)/6 --> 67.2F Jacksonville: 76.7 (76.4+76.3+75.2+73.2+74.0+80.2)/6 --> 75.9F Detroit: 58.7F (58.0+61.1+54.1+53.2+57.0+59.5)/6 --> 57.2F Chicago: 60.0F (58.3+62.0+55.8+54.4+54.5+57.9)/6 --> 57.2F St. Louis: 65.7F (63.4+64.7+60.0+61.0+62.8+63.5)/6 --> 62.6F Billings: 50.8F (46.4+51.2+33.7+54.0+42.0+48.3)/6 -->45.9F Albuquerque: 64.8F (57.1+63.4+62.0+60.0+57.7+57.6)/6 --> 59.6F San Diego: 73.3F (68.1+69.7+68.9+71.0+66.6+73.5)/6 --> 69.6F Seattle: 52.7F (46.5+52.0+48.2+55.8+49.4+51.4)/6 --> 50.6F The ten city average error jumped from 1.6 degrees to 2.2 degrees - pretty big increase for such a short period.