Griteater's Winter Outlook (17-18)

[griteater]

Griteater’s Winter Outlook (17-18)

 

Forecast Maps

 

 

 

 

 

  

 

 

“…it is our view that doing the science and going through the steps is as important as coming up with the right answer.  A lot of people don't understand this, as the emphasis in our culture is on getting the right answer.  For example, if I forecast the winter pattern based upon the number of socks in my dresser drawer and the forecast turns out to be correct, it doesn't mean my technique was scientifically based and/or it can be used in the future.”

 

Dave Tolleris (wxrisk.com) – November 2017

 

 

Forecast Parameters

 

1) ENSO – in my view, “Weak Cool ENSO” is the best way to characterize ENSO for the upcoming winter.  The average of the IRI ENSO model suite shows an official Weak Nina for late fall thru winter (total of 25 dynamical and statistical models).

 

All of the following parameters are currently in a state that is consistent with Cool ENSO / La Nina:

• VP – 120-day VP shows uplift over the Maritime Continent, centered along 120-130E, with subsidence over South America.
• OLR – 120-day negative OLR in the Maritime Continent / positive OLR near the dateline
• Global AAM – July 1 to Oct 31 daily averaged AAM is -0.50 (it was -0.26 for same dates in 2016)
• Equatorial SOI for Jul-Sep is 0.4 (also 0.4 for same dates in 2016)
• Equatorial Upper Ocean Heat Content east of the dateline is cooler than normal
• Equatorial Sea Level Anomalies east of the dateline are below normal
• MEI eased into Weak Nina territory in its October release.

 

2) Cool ENSO / La Nina Base – the Sep-Oct SST anomaly pattern is characteristic of an East Pacific Cool ENSO event.

 

I have high confidence that this will continue to be an East Pacific event.  If it were going to be a Central Pacific event, the max cool anomalies would already be located in the Central Pacific.  Image below from this paper from Zhang et al.

 

 

 

3) PDO – The PDO has dropped from its positive values of the past few years.  The PDO isn’t currently in a strongly positive or strongly negative regime.

 

4) QBO – the QBO is negative at 30mb and is slowly going through a positive to negative phase transition in the lower stratosphere (40-50mb).

 

5) Solar Parameters

• Solar Flux – solar flux has dropped into the range that is considered “solar minimum” (below 105).  Outside of a strong flare up in September, solar flux has been weak for the entire year (90 day averages have been in the 70-90 range).  Odds favor it remaining weak overall through winter.

 

• Solar Geomagnetic Activity – looking at the Ap Index data since 1933, the range for the highest 30% of Jan-Sep averaged index values runs from 17.0 to 23.4; and the range for the lowest 30% of Jan-Sep averaged index values runs from 4.3 to 11.3.  The Ap Index spiked high in September this year, but has otherwise been below long- term averages.  The Ap Index averaged for Jan-Sep this year is 10.7, within the lowest 30% of index values.

 

• Solar Wind – for solar wind data since 1964, the range for the highest 30% of Jan-Sep averaged values runs from 460.9 to 543.2; and the range for the lowest 30% of Jan-Sep averaged values runs from 371.0 to 422.6.  The solar wind for Jan-Sep this year is 460.9, just edging into the highest 30% of values.

 

 

Forecast Discussion

  

North Pacific Pattern

 

1) When the Jul to Oct AAM averages negative in concert with Cool ENSO conditions (Negative Neutral & La Nina), a 500mb high pressure ridge is strongly favored to exist in the North Pacific in the subsequent mean winter pattern (26 of 27 cases via AAM data back to 1958).

 

2) The location and configuration of the Cool ENSO North Pacific Ridge has a large influence on the weather pattern over North America. 

 

For the Cool ENSO winters that contained a North Pacific Ridge in the mean winter pattern (34 of 38 cases since 1949), I binned the winters based on the relative position of the ridge, i.e. whether the ridge was located in the ‘relative’ NW, NE, SW, or SE portion of the North Pacific.

 

 

 

 

 

 

3) The key takeaways from those images:

• The “North” North Pacific Ridge cases allow for meridional flow out of the Arctic on the east side of the ridge, with the U.S. temperature distribution having a general cool north / warm south configuration.

 

• The “South” North Pacific Ridge cases are subject to west to east zonal flow on the north side of the ridge, cutting off the flow out of the Arctic, and have a general U.S. temperature distribution of cool west of the Rockies / warm east of the Rockies.

 

4) The next obvious question is, how do we predict the specific location of the North Pacific Ridge during Cool ENSO winters? 

 

• Anthony Masiello’s (Twitter: @antmasiello) findings from 2012 showed that “North” North Pacific Ridges are favored during La Nina / +QBO winters, while “South” North Pacific Ridges are favored during La Nina / -QBO winters.  Anthony noted that designating the winter QBO phase was heavily weighted toward the lower layers of the stratosphere (i.e. 40mb – 50mb).    

 

• The problem this winter is that assigning a QBO designation in the lower stratosphere is more difficult than normal.  In the 40mb to 50mb layer, we will have a +QBO transitioning to -QBO.  The best comparisons among Cool ENSO years for the current QBO progression on the Berlin QBO chart would be the winters of 83-84 and 00-01.  The winters of 62-63 and 81-82 were also similar, but with an earlier progression into the -QBO phase; while the winter of 11-12 was similar, but with a later progression into the -QBO phase.

 

• Neither the 83-84 nor the 00-01 winter featured a high pressure ridge in the North Pacific in a location that is typical of Cool ENSO winters.  Instead, both of these winters featured a similar mean pattern of NW Pac Trough / weak NE Pac Ridge / Central & Eastern U.S. Trough.

 

• The 62-63 and 81-82 winters both featured a “North” North Pacific Ridge with negative height anomalies across large portions of the U.S.

 

• The 11-12 winter featured a “South” North Pacific Ridge, a +AO/+NAO, and positive height anomalies across all of the U.S.

 

• When looking at Cool ENSO / -QBO winters, the 500mb pattern tends to differ based on whether the QBO is in the frontend of the -QBO cycle (i.e. either transitioning into the -QBO phase in the lower stratosphere as it will be this winter, or already in the early part of the -QBO phase in the lower stratosphere) <VS.> the backend of the -QBO cycle.  The frontend -QBO winters tend to feature a “North” North Pacific Ridge and central U.S. trough, while the backend -QBO winters tend to follow the aforementioned correlation and contain a “South” North Pacific Ridge

 

 

 

• Another factor I looked at for predicting the North Pacific Ridge location was the autumn 500mb pattern leading into winter.

 

• The Oct-Nov pattern prior to “North” North Pacific Ridge winters tended to be absent of negative height anomalies in the Bering Sea, Alaska, and NW Canada.

 

• The Oct-Nov pattern prior to “South” North Pacific Ridge winters tended to contain solid negative height anomalies in the eastern Gulf of Alaska extending up into portions of the Bering Sea, Alaska, and/or NW Canada.

 

• Based on the current Oct 500mb pattern to date, and the ensemble forecast for the first half of November, this factor is also leaning toward a “North” North Pacific Ridge projection for winter since it doesn’t look like we are going to see solid negative anomalies in the mean pattern that extend from the Eastern Gulf of Alaska into the Bering Sea, Alaska, and/or NW Canada.  The forecasted pattern for the first half of November favors substantial ridging in Alaska and the Bering Sea.

 

• Forecasting the West vs. East placement of the North Pacific Ridge

 

• I looked at autumn tropical OLR and VP patterns during Cool ENSO years, but couldn’t ascertain any clear indicators to help with formulating a forecast for the west vs. east placement of the North Pacific Ridge.

 

• Conventional thought may be to assume that a Central Pacific Nina favors a “West” North Pacific Ridge and that an East Pacific Nina favors an “East” North Pacific Ridge.  However, for the Central Pacific Ninas, I did not see the data to support this notion as there was an even mix.  For the East Pacific Ninas (like we will have this winter), maybe surprisingly, the opposite was true, as “West” North Pacific Ridges were more favored (11 of the last 14 cases).  In addition, when restricting this to East Pacific Cool ENSO winters that yielded a “North” North Pacific Ridge, 7 of the 8 cases produced a “North” North Pacific Ridge that was biased to the “West.”

 

• Finally, when looking at autumn 500mb patterns, wintertime “Northeast” North Pacific Ridges tended to feature positive height anomalies in the Eastern Gulf of Alaska and along the Canadian and U.S. west coast during months of Oct-Nov.  Since noteworthy positive anomalies are lacking in this region this autumn, this factor is also leaning toward a “West” North Pacific Ridge placement. 

 

 

Bottom Line: Based on the combination of an East Pacific Cool ENSO event, the frontend of the -QBO phase, and the current & projected 500mb pattern this autumn, I favor the development of a “North” North Pacific Ridge that is variable, but biased to the northwest in the mean winter pattern.      

 

 

 

AO / NAO

 

1) When looking at AO/NAO data, I have recently preferred viewing these indexes as a lump sum average.  So, it’s simply AO+NAO/2, whether it be for a single day, month, or a full winter season.  We have been on quite a run of positive AO/NAO index values.  Over the past 4 winters (Dec-Mar), only 2 of 16 months featured a AO/NAO Avg that was -0.50 or lower (Jan 2014 & Jan 2016).

 

2) Cool ENSO is no stranger to -AO/NAO winters.  Of the 25 winters in which the AO/NAO was -0.50 or lower, 14 were Cool ENSO, while 11 were Warm ENSO.  Of note, however, is that strongly negative AO/NAO winters were more likely to occur during Warm ENSO than Cool ENSO.

 

3) Below are tables I put together containing select ENSO, QBO, Solar, SSW, and AO+NAO data.  The selected years are from portions of the last 6 solar cycles.  The first year in each of the 6 table images is close to the max of the solar cycle (flux maximum), while the last year in each table image is the last year of the subsequent minimum (flux minimum).  Here’s an explanation of each column:

 

• Winter Ending – e.g. “1959” equals the winter of 58-59.

 

• ENSO SON-JFM – winter ENSO designation based on Eric Webb’s (@webberweather) ENSO Ensemble ONI using an average of the 5 tri-monthlies from Sep to Mar.

 

• QBO (Jan-Feb Avg at 45mb) – Positive or Negative QBO designation and trend at 45mb, averaged over Jan-Feb.  45mb is used in a couple of prominent papers on Solar Forcing, the QBO, and SSWs.

 

• Solar Flux (Dec-Feb) – monthly averaged solar flux for the winter (Dec-Feb).

 

• Solar Flux Level (Dec-Feb) – max, min, or neutral solar flux designation.  150 and above = max, 105-150 = neutral, 105 and below = min.

 

• Solar Ap (Jan-Sep Prior) – monthly averaged Solar Ap Index for the Jan-Sep period prior to the winter.  Orange = Top 15^th percentile of values, for the Jan-Sep period, for Solar Ap data since 1934.  Pink = Top 15^th to 30^th percentile.  Dark Blue = Bottom 15^th percentile.  Light Blue = Bottom 15^th to 30^th percentile.

 

• Solar Ap (Nov-Feb) – monthly averaged Solar Ap Index for the winter (Nov-Feb).  Orange = Top 15^th percentile of values, for the Nov-Feb period, for Solar Ap data since 1934.  Pink = Top 15^th to 30^th percentile.  Dark Blue = Bottom 15^th percentile.  Light Blue = Bottom 15^th to 30^th percentile.

 

• Solar Wind (Jan-Sep Prior) – monthly averaged Solar Wind values for the Jan-Sep period prior to the winter.  Orange = Top 15^th percentile of values, for the Jan-Sep period, for Solar Wind data since 1966.  Pink = Top 15^th to 30^th percentile.  Dark Blue = Bottom 15^th percentile.  Light Blue = Bottom 15^th to 30^th percentile.

 

• Solar Wind (Nov-Feb) – monthly averaged Solar Wind values for the winter (Nov-Feb).  Orange = Top 15^th percentile of values, for the Nov-Feb period, for Solar Wind data since 1966.  Pink = Top 15^th to 30^th percentile.  Dark Blue = Bottom 15^th percentile.  Light Blue = Bottom 15^th to 30^th percentile.

 

• SSW – indicates if and when a SSW (Sudden Stratospheric Warming) occurred during the winter.  Data from Amy Butler’s (@DrAHButler) table on SSWs.

 

• AO+NAO Avg – AO+NAO/2 for the winter (DJFM), and each individual winter month.

 

 

 

 

 

 

 

 

 

4) The tables show the known lag with respect to solar forcing parameters thru each cycle.  That is, as solar flux comes off solar maximum and declines toward solar minimum, the geomagnetic parameters (Solar Ap / Solar Wind) tend to remain or become elevated.  Typically, the geomagnetic parameters don’t begin to reach minimum values until the end of the solar flux minimum period.

 

5) The yellow highlighted years are the 5 years with the combination of: 1) the frontend of the -QBO (either Neutral & Falling or Negative & Falling), and 2) solar flux minimum.  2018 is also highlighted as the anticipation is that we will add a 6^th year to that list this winter.  For the 5 highlighted winters, 16 of 20 months had a -AO/NAO, the highest percentage for any QBO/Solar Flux combo.  The stars aligned during the 09-10 winter as all solar parameters were very low in concert with a frontend -QBO. 

 

Note: for 95-96, the QBO wasn’t as far into the negative phase as the other highlighted years, but I like to include it in the frontend -QBO group because the QBO was already solidly negative at 30mb at the beginning of 1996.

 

6) The 76-77 and 86-87 winters appear to be the best QBO/Solar comparisons for this winter.  Leading up and into those winters, solar flux was low, the Solar Ap index was neutral or slightly low, and the solar wind was neutral or slightly elevated.

 

7) As long as we don’t see another significant spike in solar activity like we did in September, in my view, the current QBO/Solar environment is one that is certainly more favorable for -AO/NAO periods than we’ve seen in the past 4 winters.  If we do see a spike in solar activity, that would somewhat negate those ideas.

 

8) With respect to the early developing stratospheric polar vortex, we aren’t seeing indications of a super strong or super weak vortex as of early November.  Based on the October conditions and what is forecasted for early November in terms of zonal wind anomalies and heat fluxes, I would call it slightly stronger than normal.  Image from Hannah Attard (@HannahAttard)

 

 

 

9) Some numbers/thoughts regarding SSWs:

 

• 4 of the 5 winters with the combo of frontend -QBO and solar flux minimum yielded SSWs.  Limiting those 5 winters to the 2 that were Cool ENSO, 1 of 2 had a SSW.

 

• When including all -QBO winters (frontend and backend) with solar flux minimum, 9 of 12 yielded SSWs.  Limiting those 12 winters to the 7 that were Cool ENSO, 4 of the 7 had a SSW.

 

• Using solar wind as the solar parameter instead of flux, in concert with -QBO, yields similar SSW numbers/ratios.

 

• The last official SSW occurred in the 12-13 winter.  I would put the chances of an official SSW occurring this winter at 50%, or equal chances (higher chance than last winter).

 

10) Image below shows the November 500mb patterns in Cool ENSO years prior to the Top 5 most negative AO/NAO winters vs. the Top 5 most positive AO/NAO winters.  Key difference is noted over the NE Pacific and Alaska.  This November appears to be aligning moreso with the image on the left, though it’s not a perfect match and we are still early in November.

 

 

 

11) Findings from the previously referenced paper from Zhang et al. showed that -NAO was favored during East Pacific La Ninas and +NAO was favored during Central Pacific La Ninas.  In doing my own research on this topic, I found that the Central Pacific to +NAO correlation was significantly more robust than the East Pacific to -NAO correlation when extending the research back to the late 1800’s, and when including Negative Neutral years.

 

 

Bottom Line: given the points made above, my forecast calls for the Dec-Mar averaged AO/NAO to be in the range of 0.00 to -0.75, or in the range from the negative side of neutral to slightly negative.  It’s my belief that we will see a few quality AO/NAO blocking episodes this winter that are more than transient in nature.

 

Verification will be using NOAA CPC Monthly AO and Hurrell PC-Based NAO.

 

 Other

1) Here is the 500mb pattern for the current year (Jan-Oct), against the 1981-2010 climatology.  As we’ve seen over the past 5 or so years, the amount of Northern Hemisphere area covered by above normal heights far outweighs the area covered by below normal heights.  Simply put, we are currently in a period of noteworthy background warming.  It’s my belief that this background warming must be taken into account with respect to temperature forecasts.

 

 

 

Winter Forecast Thoughts

 

1) December Forecast: -PNA pattern.  Western U.S. trough and Eastern U.S. ridge.  Good pattern in general for early snow at western ski resorts.

 

2) January Forecast: In my view, January holds the most promise for the development of a -EPO, -AO/NAO pattern, with possible bowl-shaped trough over the lower 48 at times.

 

3) February Forecast: There is a Cool ENSO tendency for the North Pacific Ridge to retrograde farther to the west into the Bering Sea during the month of February, opening up the door for increased southeast U.S. ridging.

 

4) March Forecast: I’m calling for a similar pattern to March last winter, but slightly cooler, with cool anomalies draped across the northern U.S and warm anomalies to the south and southwest.

 

5) As previously mentioned, It’s my belief that we will see a few quality AO/NAO blocking episodes this winter that are more than transient in nature.

 

[40/70 Benchmark]

Nice work.

I like the dichotomy that you presented between "front loaded" -QBO seasons, and "back loaded", as DT did some similar work that illustrated the fact that -QBO is more highly correlated to -NAO/AO when it remains stout throughout the winter....ie is well established to start and doesn't fade.

As a community, we have made tremendous strides in relation to analyzing atmospheric oscillations on a more profound and extensive level...where as just a few years ago we were still lumping seasons together by general QBO and ENSO phase, etc. Your outlook represents a splendid tutorial for those yearning to understand why not every like phase of an oscillation is created equally. It is this that will ascend our rather meager level of success heretofore as it relates to seasonal forecasting to the next level IMHO.

All that being said, I don't see any glaring signs of a horrible winter.

[ORH_wxman]

Very nicely written up. Good luck on your outlook. As a resident of New England, I hope it verifies.

[wolfsheepsheadbay]

Awesome work!

[griteater]

^ Appreciate the nice words all.  It means a lot.  40/70 - agree with your points that we are making progress as a group (hobbyists and mets alike) with seasonal forecasting, piece by piece.  More important to me is that we are putting in the effort.

[griteater]

I fixed a typo on the snowfall map.  It previously read "16-17."  I updated it to "17-18."  I also added plus and minus signs on the 500mb anomaly map to indicate positive and negative height anomalies.

[RUNNAWAYICEBERG]

You put a lot of time into this, broken down very well. Applause. 

GL to all winter weenies. 

[raindancewx]

I think you have the right idea. Not expecting it to be super warm out here in the actual meteorological winter - Dec, Jan, Feb, but we will torch hard in March, and so for Dec-Mar your anomaly looks right out here.

The PDO should come in right around 0 if the method I've been using to predict Nov-Apr is worth anything - so would imagine it is much drier in the SW than last year. Nino 1.2 temps and the Mar-Aug PDO predict the Nov-Apr pretty well.

Some of the years I like hint at a super cold February or March nationally outside the SW, but its not in all the analogs, so probably just powerful cold shots.

[stanleyjenkins]

Great work on the write up.  Hard to argue your forecast with all of the reasoning behind it.

How long did it take you to write up that post?

[griteater]

2 hours ago, stanleyjenkins said:

Great work on the write up.  Hard to argue your forecast with all of the reasoning behind it.

How long did it take you to write up that post?

Thank you.  For me, I'm monitoring things during the summer and into fall, then write it up a little at the time, piece by piece...some of it you have to wait until the end.

[Quixotic1]

Griteater, I like your thinking.  For your sake i hope it verifies.  For mine, I hope it doesn’t.  It would mean three torch winters in a row.  Something I’ve never seen before.  Good luck.  

[stadiumwave]

Great work!

[raindancewx]

Also, and this just my view - but I think there will be moments when the North Pacific is in all four positions you outline because the MJO is active and has some influence on it. My ranking of their duration would be something like this for Dec-Feb, with the caveat that I think the NE image briefly becomes dominant in H2 February or March. Its interesting seeing you got the patterns with location of pressure anomalies, I'm lazy - just do it with Modoki values.

1) NW (40%) (H1 Dec, H2 March, late January)

2) SE (25%) (I think this is February - first three weeks anyway)

3) SW (20%) (This is the October pattern - and the "Major Hurricanes Hitting Puerto Rico" pattern, I like it for mid-Dec to mid-Jan)

4) NE (15%) (any remaining days)

 

[Isotherm]

Very well researched, succinct, cogent analysis. Always enjoyable to read your outlooks, Grit. Your disambiguation of the various Pacific mid level ridge permutations was especially compelling. I commend you for employing unconventional, out of the box thinking, and not simply utilizing ENSO status to broad-brush (which is what almost every other outlook I've read this autumn does do).

 

[donsutherland1]

Very nice and thorough work, Griteater. Good luck with the forecast.

[griteater]

10 hours ago, Isotherm said:

Very well researched, succinct, cogent analysis. Always enjoyable to read your outlooks, Grit. Your disambiguation of the various Pacific mid level ridge permutations was especially compelling. I commend you for employing unconventional, out of the box thinking, and not simply utilizing ENSO status to broad-brush (which is what almost every other outlook I've read this autumn does do).

 

Thanks Iso for the kind words and feedback...it's appreciated.

7 hours ago, donsutherland1 said:

Very nice and thorough work, Griteater. Good luck with the forecast.

Thanks Don, good to hear from you

[OHweather]

That was a very good read.  Good luck!  Your research was very detailed and in some cases pretty unique. 

[IntenseBlizzard2014]

I'm very impressed with your analysis. That QBO/Solar Activity analysis was very enthralling.

[Syrmax]

Thanks for the work here.  Very informative and well put together / reasoned.  Hope it verifies. I'm sure it'll be interesting to see what worked and what didn't.

[raindancewx]

The Canadian is on board with something akin to what you have labelled as the "SW" N. Pacific Ridge for Dec. I'd say that is a good omen for your outlook. Fingers crossed?

Your temp anomalies look good for Dec if the Canadian is right, and similar to what I had too (https://tinyurl.com/yam55n95) although I'm colder than the Canadian nationally for December.

[WidreMann]

1 hour ago, raindancewx said:

The Canadian is on board with something akin to what you have labelled as the "SW" N. Pacific Ridge for Dec. I'd say that is a good omen for your outlook. Fingers crossed?

Your temp anomalies look good for Dec if the Canadian is right, and similar to what I had too (https://tinyurl.com/yam55n95) although I'm colder than the Canadian nationally for December.

I don't know how this is going to work with the predicted deep trough over the central and eastern US for the middle two weeks of December.