raindancewx

Monte Carlo simulation is actually a pretty good way for them to handle the MJO. It's pretty noisy as a signal. It's much simpler just to use the 60 and 70 percent composites for warm/cool, wet/dry on the MJO site. The real question I suppose is why no one reads what the CPC stuff actually means before quoting it. The placement of the greatest precipitation anomaly for winter on the Canadian is actually reminiscent of phases 6-7, both of which are warm in the East. Of course any time I say something like this, it's ignored or some random idiot on here yells at me on my Twitter for a few days. The centering of the enhanced moisture never really gets east of 180W in a given month, which is not a particularly cold match for the East on the MJO composites. The real MJO/RMM stuff tends to have little motion or conflicting patterns in stronger El Ninos, so I'm just using the general look for illustration. I don't think the actual "MJO" will be a main issue this winter. Keep in mind, this is what last year looked like on net - it was a clear phase four-five look (top right, i.e. cold West), with pronounced wetness by Indonesia sandwiched East & West by dryness in Nino 4 and the Indian Ocean. So yes, this stuff does have some skill.

A blend of 1982/2012 with 1980 taken out to amplify it was actually ~close (B ish) globally for 500 mb in July-September. If you roll that period forward, you end up with a look that is pretty similar, albeit tilted, to my analogs for winter. That said, I don't expect the fruition of this pattern to be as extreme as shown for all three months. It should be a dulled version for DJF, with only one major severe period. That's been my thinking for a while. I was pretty happy with 1984/2012 as a matching Summer blend last year, and that worked OK for winter. This year, I think 1982/1951 is a better blend for temps/precip overall, but 2012/1982 is probably better at 500 mb, and the results are similar anyway.

So now that almost the entire country is forecast to be warm in October, particularly in the East, the near exact opposite of the famous eastern years like 1965, 1976, 1977, 2002 and 2009, even 2014 and 1957 to some extent are we still pretending those are the top analogs or what? I haven't been following much, since I'm finishing up my forecast. I would have liked to see the Canadian show the Northwest colder this month per my ideas, so I'll refine them a bit more, but this is still a decent match to my unweighted blend for winter. I do think it's fairly likely this is not the right pattern for October anyway. The models kind of suck at forecasting October and March. Some of you guys must have some kind of match that's at least half decent by now - let's see them.

Still looks like a blend of strong / east based El Ninos, with 1993 subtracted out to fix the PDO. I've mentioned this on/off since August. You do have to warm up TX to account for persistence and the prior dryness. We've had this pattern on/off since June at this point, parts of the East/West normal/cool with the middle of the US warm, and occasional flips to the exact opposite. If that doesn't reset in October, there is no reason to expect it to not continue. The El Nino has weakened below the surface again, which has tied in recently to warmth in TX and the states near it. When we had the warming subsurface, you had a major cold shot in the Plains mid month. If you try a weaker blend, like 1994/2004/2006 on the MEI or RONI, you end up missing the heat and the coolness.

The patterns globally have actually been very El Nino. None of you actually are interested in looking at ENSO surface patterns before the winter because you can post stupid upper level and forcing maps that have virtually no correlation to ground conditions in Summer. There is literally no correlation between PNA and Nino 3.4 in the pre-winter months. 99% of you take El Nino / La Nina as nothing except a PNA sign indicator. If anything August has a negative correlation to Nino 3.4, so you are supposed to have -PNA signatures as we've seen at times in the Summer with an El Nino. PNA is such a weak signature in particular in June-July that CPC doesn't even calculate it. If you're too lazy to know the actual patterns then of course you'll continue looking for signatures that don't exist.

Looking back, there are actually quite a few stronger El Ninos with active hurricane seasons in September. Another silly example of forcing not being used correctly to me. We're at about 65 ACE, and it will grow somewhat more with the storm to hit the Carolinas. A bit ahead of the most active El Ninos in the previous warm AMO cycle. But not dramatically. Most of you seem wedded to the idea that years like 1965 and 1957 were well coupled El Ninos, but of course, the 30-year average ACE in September for 1935-64 is something like 44. The 1991-2020 average is more like 60-ish. So we're not even as far above the recent average as those older strong El Ninos like 1957 and 1965 that I think everyone agrees are well coupled in other ways. A lot of El Ninos actually have pretty impressive early season hurricanes and total activity, call it June-September, and then die prematurely in October from what I can see. https://psl.noaa.gov/gcos_wgsp/Timeseries/Hurricane/hurr.atl.ace.data 1951: 57.8 ACE in September 1953: 58.6 ACE in September 1957: 63.7 ACE in September 1965: 60.9 ACE in September 2002: 46.8 ACE in September 2003: 111.1 ACE in September (if considered an El Nino) 2004: 155.0 ACE in September 2006: 59.6 ACE in September 2018: 72.8 ACE in September 2019: 93.4 ACE in September

Here is what BOM says about declaring El Nino. For my purposes, this El Nino started in March, and is already half over. So it's actually peaking at the surface between now and 11/1. Subsurface will have fits and starts through Oct/Nov before steadily weakening. Unlike winter, the greatest Spring periods for severe cold and precipitation in the West are during the most rapid ENSO transitions. You can look back at Spring 1973, 1988, 2005, 2017, 2019, 2023, etc to see this effect. The joke is the Summer temp profile was a pretty strong match to actual Nino 3.4 SST correlations in the Pacific, despite the constant bitching about how there is no forcing. Who cares? If you're getting El Nino conditions in the expected way, does it matter if not all the signals show up as you'd expect? Given how weak the correlation is, in theory it takes remarkably warm SSTs to drive such a good match - which of course we had in June-August. El Niño and positive Indian Ocean Dipole underway An El Niño and a positive IOD are underway. The declaration of these events, and their concurrence over spring, reinforces the Bureau's long-range rainfall and temperature forecasts, which continue to predict warmer and drier conditions for much of Australia over the next three months. The confirmation of an established El Niño increases the likelihood that the event will be sustained through the summer period. Oceanic indicators firmly exhibit an El Niño state. Central and eastern Pacific sea surface temperatures (SSTs) continue to exceed El Niño thresholds. Models indicate further warming of the central to eastern Pacific is likely. Broadscale pressure patterns over the tropical Pacific reflect El Niño, with the 90-day Southern Oscillation Index (SOI) at −7.7. Recent trade wind strength has been generally close to average, but was slightly weaker than average across the tropical Pacific in August 2023 for the first time since January 2020. Overall, there are signs that the atmosphere is responding to the pattern of SSTs in the tropical Pacific and coupling of the ocean and atmosphere has started to occur. This coupling is a characteristic of an El Niño event and is what strengthens and sustains an event for an extended period. Climate models indicate this El Niño is likely to persist until at least the end of February. El Niño typically leads to reduced spring and early summer rainfall for eastern Australia, and warmer days for the southern two-thirds of the country. A positive Indian Ocean Dipole is underway. The Indian Ocean Dipole (IOD) index is +1.25 °C for week ending 17 September. This is its fifth week above the positive IOD threshold (+0.40 °C). The longevity of this trend, combined with the strength of the dipole being observed and forecast, indicate a positive IOD event is underway. All models predict this positive IOD will persist to at least the end of spring. A positive IOD typically leads to reduced spring rainfall for central and south-east Australia. When a positive IOD and El Niño occur together, their drying effect is typically stronger and more widespread across Australia.

My two main analogs for the winter are 1982 and 1951 at this point. 1951 had a major mid-continent, mid-September cold shot. Not exact, but it's a -PDO, east based El Nino...it's more the tilt of the pattern that is off, v. the actual indices per se.

This time of year, the PNA has no actual correlation to Nino 3.4. But the PDO does, and so I do think I'm on the right track going with the PDO as the temperature predictor for the winter and the ENSO as the precip predictor. This is my tentative blend for the winter. It's not super different than the Canadian, although obviously more blue south of Alaska. The out-of-season heat waves in Mexico in developing El Ninos tend to precede very cold periods there. The hot season on the plateau is more like March-May. The heat there means the monsoon process/subtropical jet/moisture etc was moved from its normal latitude to the north or south, allowing dry air to work in, instead of the rainy season. The corresponding movement in the winter allows cold dry air to accompany much colder lows on the plateau and it gets very cold. Years like 1982 have monthly record heat in June, and then severe cold in winter. Again...volcanic Summer in Mexico - 1982 was too. The five hot "developing El Nino" Junes on the Mexican highlands - Most El Ninos are cold in June in Mexico - so it's been interesting looking at those five as a hint for what the subtropical ridges will do.

The Navy guy already linked it a month or so ago, but there is a pretty good paper out about volcanic effects from Tonga. Before this paper, I've mentioned that volcanoes tend to screw up the placement of the ITCZ which has lots of implications for weather outside the tropics. https://essopenarchive.org/users/304243/articles/657090-long-term-surface-impact-of-hunga-tonga-hunga-ha-apai-like-stratospheric-water-vapor-injection I've been saying for ages now that March is favored for very powerful storms in the West with both high solar and volcanic activity. You can see the little blue area for precipitation in the paper for my area. Now look at JJA - it's not a horrible match for the Summer in North America. It was very hot over the Mexican plateau - El Ninos are correlated to warmth in Mexico but it tends to be strongest as a signal on the West Coast there, not inland and high up. The East had a fairly cool Summer. You can see the tendency is to flood Canada with warm air for winter, while Kamchatka / Japan remain cool in winter. That's maybe a pseudo +PNA v. displaced +WPO type look. Long-term impact surface temperature and precipitation anomalies forced by the stratospheric water vapor cloud from Hunga Tonga - Hunga Ha'apai.

PDO is still extremely negative. https://oceanview.pfeg.noaa.gov/erddap/tabledap/cciea_OC_PDO.htmlTable?time,PDO 2023-08-01T00:00:00Z -1.68 El Nino, with -PDO values in August on this scale. 1951, 1953, 1963, 1969, 1991, 1994, 2006 I've been seeing people say this September is like 2004 because of the hurricane activity in the Atlantic / weak El Nino sense. In a sensible weather look though those Septembers have been nothing alike so far. We're actually at like B- similarity to early September 2015, which also had a big hurricane Bastardi and some of the other famous dudes thought would hit the East Coast. I think it was Isaias? (Edit: It was Joaquin). We certainly are not similar to 2009 at the moment. The cold in the West is wrong by placement in 2015 v. 2023 but you at least have a warm streak in the Plains and Northeast like this year.

This Summer in some ways reminds me of 2016. That was a winter that had Nino 1.2 warm relatively, with the western areas somewhat cool. So you had a strong subtropical jet / western storminess in winter, following a very hot Summer. The global upper level pattern was different in the Summer, especially on the Atlantic side, but you could do a lot worse for a 500 mb match than combing something like 1982, 2016, 2019 for a match to 2023. I'm not a fan of the QBO as a relevant metric, but the -QBO years do seem to be better matches right now. 1951, 1972, 2009, 2019, and a few others. The bigger El Ninos have tended to be near average QBO readings - 1997, 1982. The 2015-16 El Nino got blamed for breaking the QBO from what I remember, in terms of the regularly scheduled timing. The joke is the stuff that is determined by actual SSTs, like precipitation - looks very similar to the major El Ninos. It's just that the temp patterns are not great matches. The temps are more correlated to the PDO though in terms of spatial layout in the US. You can bitch and moan about the influence not behaving like you want - but this is pretty close. The upper high over TX was stronger than in the matching years, and CA had remnants from a hurricane...but its not like it's completely opposite the major El Ninos or something.

-PDO doesn't favor snow in the Mid Atlantic in El Nino. I said no correlation, not positive. They get so little snow that their good years are more fluky, even in El Ninos or favorable periods. But it is a negative feature for areas in the Northeast with more snow. In non-El Nino years, -PDO is bad for all of you, since it usually comes with La Ninas. Typically it is bad for the Southwest as well. The -PDO years favor warmth in the East and dryness in the Southwest deserts.

I looked at all the "counter" cyclical winters for PDO v. ENSO today, using Nov-Apr for the PDO. I'm using the old PDO index, not the stupid normalized index that replaced it. So 2009-10 as an example comes in slightly positive for Nov-Apr. Anyway, these are the El Ninos with a -PDO or a La Nina with a +PDO. Just about all of them have a major "snow hole" in the Northeast, which I'm defining as Virginia and north. When I say snowless/low snow, I don't mean no snow. Just that the entire region is below average in a notable way, despite surrounding areas or neighboring areas of the US/Canada being quite snowy. I'm basing these on snow maps for July-June from the MRCC site. The counter cyclical ENSO/PDO years seem to have more consistent storm tracks. So the Northern stream stuff whacks the same areas over and over and the southern stream stuff whacks areas over and over, and those tracks don't really vary much. A lot of other years have much less consistent tracks, and so you get less spread in the anomalies whether high or low for a given spot. https://mrcc.purdue.edu/CLIMATE/welcome.jsp 1933-34 - interior NE snowless 1938-39 - interior Virgina snowless 1942-43 - New England snowless 1983-84 - NJ / CT / E. PA snowless 1984-85 - Snowless outside W. NY 1985-86 - snowless outside W. NY 1995-96 - very snowy 2000-01 - snowless PA/VA 2005-06 - snowless NE US 2017-18 - snowles central VA/DC 1951-52 - low snow NJ, e. PA, W. NY 1953-54 - snowless 1958-59 - low snow NJ, E. PA, CT 1963-64 - low snow Maine, W. NY 1965-66 - low snow NY/CY/N PA 1968-69 - low snow PA, NJ, VA, W. NY 1972-73 - low snow PA/NJ/Southern NY & New England 1994-95 - snowless 2006-07 - low snow outside W. NY/PA 2019-20 - low snow outside Maine

For Boston, if you graph all El Ninos out by the PDO value for Nov-Apr, there is a pretty clear signal for the -PDO El Ninos to see less snow. That signal vanishes completely by the time you get to Philadelphia, and we also have no PDO snow signal in El Nino locally. Every blend I've been able to come up with has near normal snow for most of the Northeast US, outside of Southern New England, which is generally consistent with the image below.

Apparently, it was in December though? It's not like the warm pool wasn't there in December. You can predict whatever you want. But I'll take my chances on the PDO given that the western warm pool will continue to weaken relative to norms while the ENSO will be most pronounced relative to norms in that time frame.

There's never really been any correlation between the PNA and Nino 3.4 in December. It's best to treat ENSO as something that can, but doesn't always influence the various indices. The PDO on the other hand, is very strongly correlated to the PNA, much more so than ENSO in Dec-Jan. For Feb/Mar the PDO/ENSO essentially tie. I'm very much more on board with riding the PDO early for temps, and then dulling canonical El Nino effects later if you still have -PDO / +ENSO in Feb-Mar. https://psl.noaa.gov/data/correlation/table/corr.table_dec.txt

Locally, I can't find any 12-month periods with meaningfully less than ~4 inches of rain back to 1892. Our local "wet" season of July-Sept may finish under 1-2 inches at this point (v. 4" as average). El Nino precipitation long-term is about 10-30% above average by season. But even average precipitation on an annual basis would require a massive shift, since we only average 0.5"/month from Oct-May. My point with this is that US precip patterns in Jun-Aug look similar to a lot of the bigger El Ninos. So the subtropical jet / Pacific farting up moisture aspect of the El Nino looks correct, even if it is just from enhanced ocean heat rather than the canonical air pressure patterns. For June-Aug, you have pretty diametrically opposite correlations to the -PDO and +ENSO. So I don't find it surprising that some areas are responding to the -PDO and others to the +ENSO pattern. The -PDO is actually more correlated to the response by Indonesia than ENSO is in Summer.

This is my original map of the idea with site specific snow totals. So less snow as the yellow zone moves north.

I think the 60F annual temperature threshold is probably somewhere near Richmond in real time tbh. Just to be clear - the map I made is 7.25C to 18.25C in increments of 2.75C. So it's not exactly 5F, but very close. Part of the issue you guys have is that it's rarely cold enough for good snow outside of core winter, like 12/15-3/15 roughly. Here in the desert, it's certainly warming at night. But it still gets really fucking cold sometimes in Oct-Nov, and Apr-May. So we get fairly regular fluky snow in those months. You guys don't really have that to fall back on once the core season starts to warm/shrink. In the past five years alone I've had lows in the 30s every month from September-May. My point is the increase in moisture is more meaningful when its spread out over a longer zone of opportunity for snow. But you guys have a tight window for snow. Not like here when we've had snow over a foot of snow in Oct-Nov, Mar-Apr just in the past five years.

From about 20-50N, this is what I've found in terms of snow, when mapping out several hundred locations by their annual temperature v snow. In colder areas in the tropics (high mountains) it still doesn't snow as there is too little variation from Summer to Winter temps. <45F: Heavy snow is possible most of the year. Sites average 40"-750" (Bismarck, Caribou) 45-50F Heavy snow is is common in Fall-Spring. Sites average 30"-175" (Denver, Flagstaff, Buffalo) 50-55F: Snow is common, and frequently heavy in winter. Sites average 10"-60" (Boston, Santa Fe) 55-60F: Snow is common in winter. Sites average 3"-30" (Albuquerque, Philadelphia, Parral) 60-65F: Snow is uncommon, but not unheard of. Sites average 1-10" (Atlanta, Birmingham, Chihuahua city, Durango city) >65F: Snow is rare, sites average 0-3" (Houston, Miami) I've not really found exceptions at 60-year time scales for annual temps v. the snow ranges I'm listing above. Generally, the more arid spots are on the lower end of the scale, while sites adjacent to large bodies of water or on mountain tops are at the high end. Parts of the Northeast / Mid Atlantic (NJ / PA / MD / DE / VA) should be moving into the 60-65F zone pretty soon, if they aren't there already. It's still pretty common to get snow 3-7 winters a decade in those zones. Even down to old MX, Chihuahua (mean temps are typically w/in 2F of 65F) still gets ~3" on average with snow in ~half of winters. Albuquerque has had snow every cold season on record, at an average temperature of about 59F annually, but some of those years are under 1.0" and sites just 20 miles to the south or 300-600 feet lower in elevation go 1-3 years with no snow. The white xs in Mexico are examples of cooler towns that still semi-regularly see snow. The point is to show how it's tied to temps and not latitude. The green zone sees snow in ~nearly all or all winters. The yellow zone is where you can go years without meaningful snow. You guys are on the border of the yellow / green zone now, with the yellow advancing north. What I've found near the transition from the 55-60F zone to the 60-65F zone is that big March storms tend to start happening more frequently in mid-late February, with March starting to behave more like a full-on Spring month.

CPC updated Nino 3.4. August came in at 28.16C, which they call +1.30C. In older periods, when the 30-year average is lower, it's more like +1.50C. You can see the SSTs are for August in Nino 3.4 are colder than 1987, 1997, 2015. Actual SSTs are pretty close to a blend of 1997 / 1972 in August outside Nino 4. Year: 1.2 / 3.0 / 3.4 / 4.0 1972: 23.38 / 26.74 / 27.94 / 28.75 1997: 24.96 / 27.71 / 28.74 / 29.29 Blend: 24.17 / 27.22 / 28.34 / 29.02 2023: 24.33 / 27.09 / 28.16 / 29.69 Not really sure why everyone is obsessed with the Indian Ocean / Nino 4 being warmer than the Eastern zones - you can clearly see that was the case in years like 1997 too.

It's actually pretty amazing how similar US temps for Jun-Aug overall are to Jun-Aug 1982, if you add 1F to 1982. That's despite all the bitching about not locking in / getting the forcing we're supposed to get. I mean...1982 had the forcing right? I posted some maps in the winter 2023-24 thread for anyone curious.

Here is 1982 warmed up 1F v. 2023 for Summer - The pattern I've noticed in looking at three month periods is we follow 1982 pretty well for the placement of subtropical highs and heat to the south, but not so much the placement of the cold air. I'm building my analogs around the idea that the stronger El Ninos are probably a better match for TX/MX and the SW US, an OK match for areas just north/east of that zone, and not a good match for most of the Northern US. Actually think Northern Mexico may have a really cold winter, while no one in the US is materially below average. -PDO is a pretty strong late winter warm signal for the East, but a cold signal for the MX/ US West. El Nino is a strong cold signal SW US / MX in December, and January would be a blend. Last year I liked 2012/1984 as the primary temperature blend idea for winter, with the other years mostly in there for precipitation and timing fixes. This year I like 1951/1982 as a blend, with 1972/2009 likely to show up as the El Nino weakens in late Winter or Spring, and 1997/1991 in there for early winter. I do think 1972, 1982, 1997, 2009 are the better analogs for precipitation. See for yourself - Feb - Apr March-May

I don't think it's physically impossible for Nino 3.4 to gain something like 4C year over year. But it's never come close to doing it. In close to 100 years of data even a 2C gain is very rare. Last winter was 25.95C in winter. Nino 3.4 averages ~26.5C in winter. No reason to believe this will go much above 28.0C for winter. Even those 2.0C gains are like once a decade. May is also a good indicator - you don't really warm up much in DJF from May even in super strong events.