Futility Thread - Winter 18/19

[LibertyBell]

On 2/2/2019 at 10:38 AM, 40/70 Benchmark said:

This is how most people think, and why most laymen don't attempt long range forecasting.

Honestly though, meteorology isn't an exact science like, say, astronomy is- there is an art to recognizing patterns (which is a talent I see you have.)  It will never be an exact science and be able to make perfect predictions like astronomy can for eclipses hundreds of years from now, for example.  Thats why NOAA does probability outlooks.

 

[LibertyBell]

On 2/2/2019 at 10:50 AM, 40/70 Benchmark said:

That super el nino was not as horrendously east-based as 1997 and 1982....which is why I nailed the mid atl blizzard. I just went overboard and forested too much snow in NNE, and it was awful up there.

I know 1983 had the megalopolis deal, but I think odds are elevated when it isn't so extremely east based.

1982-83 and 2015-16 were amazing for what we had to work with.  And the Jan 2016 HECS proved that Feb 83 wasn't a one off deal- it was 1997-98 that was the outlier because it didn't have an HECS in our region.  2015-16 had a decent February too and a below zero temp in NYC on Valentine's Day!

 

 

[LibertyBell]

On 2/2/2019 at 10:48 AM, CoastalWx said:

If you want to talk about sort doing ok in an awful pattern, 15-16 is it. I think my total was 40.9” with a 10.5” storm to boot. All amid an all time torch winter. How ‘bout them apples?

2015-16 was a lot like 1982-83 and we've recently seen a trend to big precip events..... we had 40" of snow that season with a 40 degree average winter temp!

 

[LibertyBell]

On 2/2/2019 at 10:58 AM, 40/70 Benchmark said:

Well some of it is explainable....with no blocking, there is an avenue for every storm to cut, and with poorly timed confluence, also an avenue for every storm to whiff south.

But when all of them avail themselves of said opportunities? I call it SNE's Patrick Mahomes impersonation at the meteorological coin flip.

Imagine if this happens for the next 10 years and then you'll understand the pain of the 80s lol.  There may be a factor of luck/variance but the fact is that these kinds of patterns are very stable and once they set in, are very hard to dislodge.  That coin you mentioned would be a loaded one.

 

[LibertyBell]

On 2/2/2019 at 10:54 AM, WinterWolf said:

Exactly what we’re twlking about.  Nobody can put a finger on what made that work out ok...and this season sucking ass.   

It worked out because the chances of an HECS were relatively high in a nino that strong.

[raindancewx]

Looking at highs in Boston for 12/1-2/18, it looks like 1994-95 is within 0.2F for 90% of the winter, so I'm fairly happy with that as the analog I double weighted for this winter. It is within 0.1F for average temperature.

[CoastalWx]

6 hours ago, raindancewx said:

Looking at highs in Boston for 12/1-2/18, it looks like 1994-95 is within 0.2F for 90% of the winter, so I'm fairly happy with that as the analog I double weighted for this winter. It is within 0.1F for average temperature.

Except how it happened was not like 94-95. We had a lot of SE ridging. 

[ORH_wxman]

31 minutes ago, CoastalWx said:

Except how it happened was not like 94-95. We had a lot of SE ridging. 

Specs aren't close either. Dec/Jan '94-'95 was so much warmer than this year. Feb '95 was frigid unlike this year. 

[40/70 Benchmark]

Let me preface this by saying that Raindance has had the best outlook thus far and mine has been bad. However given that this season's la nina like forcing regime has eaten the consensus alive, I'm not sure that I'd be extolling the double weighting of any el nino analog as a virtue of any outlook.

I'll leave it at that.

[EastonSN+]

Even though I have received more snow then some past years, so far this is the worst winter in my lifetime. At least those years were nice and warm.

Hopefully March saves the day.

[raindancewx]

My main misses for the winter will be the South and Northern Plains for temps. I think Bismarck was +7 in December but will finish -2 or -3 for the winter which is pretty remarkable. I've been pretty happy with my snow map which had the Plains and Rockies doing very well, and also upstate NY.

I don't actually care about upper air patterns too much - but this was my preliminary look at my outlook for highs a few days ago. I have some theories on why the SOI went so positive in December that I'll be looking for in future El Ninos.

DJF High (F) Verification (12/1-2/16)	1953	1976	1986	1994	1994	2006	Mean	Actual	Error
Atlanta	55.4	45.9	53.0	55.7	55.7	56.1	53.6	56.9	3.3
Albuquerque	49.8	46.4	47.2	52.5	52.5	44.9	48.9	47.9	-1.0
Amarillo	54.4	49.1	49.3	53.7	53.7	45.2	50.9	55.0	4.1
Billings	38.6	37.1	43.4	39.7	39.7	35.8	39.1	34.6	-4.5
Bismarck	27.0	21.1	34.2	22.8	22.8	26.7	25.8	23.1	-2.7
Boston	40.4	33.9	37.9	41.3	41.3	40.7	39.3	41.8	2.6
Denver	50.2	48.3	46.3	47.9	47.9	36.9	46.3	45.6	-0.6
El Paso	59.4	57.4	57.9	61.9	61.9	56.2	59.1	59.3	0.2
Jacksonville	67.3	59.2	65.1	65.4	65.4	68.2	65.1	68.4	3.3
Philadelphia	44.3	34.1	41.4	44.8	44.8	44.6	42.3	44.2	1.9
San Diego	67.1	70.6	66.1	64.1	64.1	64.3	66.1	65.9	-0.1
Seattle	44.9	49.1	48.4	49.3	49.3	45.9	47.8	48.0	0.2
St. Louis	46.6	33.5	42.1	41.3	41.3	42.0	41.1	43.2	2.1

[40/70 Benchmark]

Interesting...well, that explains alot. Tricky proposition forecasting sensible weather outcomes with blatant disregard for upper air patterns. I don't think I'll venture to try that.

[weathafella]

Well BOS is up to st least 10.5 now.  

[NorEastermass128]

20 minutes ago, weathafella said:

Well BOS is up to st least 10.5 now.  

Now that futility is over, let’s shoot for normal!  

[LibertyBell]

On 2/20/2019 at 7:17 AM, ORH_wxman said:

Specs aren't close either. Dec/Jan '94-'95 was so much warmer than this year. Feb '95 was frigid unlike this year. 

We've had a strong SE Ridge this winter which is why the SE has been baking (Atlanta had its earliest 80 ever) and we've been in a mean west trough east ridge pattern.  The SE Ridge is why the mean storm track has been unfavorable for coastal areas.  94-95 isn't a good analog because we've had zero benchmark coastals this year- it's much more like 07-08.

We haven't been anywhere close to a decent snowfall around here, maybe different up north by you, but the reason why I dont say bad luck played much of a factor is because no storm came anywhere close to being a significant snowstorm around here.  I dont place much weight in average monthly temps because you can easily have a winter pattern where it's normal or below normal temps 6 days out of the week and above normal temps on the day the storm comes in- because temps on storm day are defined by storm track.  And it doesn't need to be very cold to snow, having it just cold enough is actually better to maximize moisture content and not too much which would create a suppressed pattern.  High winds make sure the snow ratio stays close to 10:1 anyway.  We've had many winters when it was cold like in the 80s with January avg temps in the mid 20s and we had a lot of precip but not much snow because the pattern dictated a bad storm track.  The pattern makes the snowfall more than average temperatures do.

[LibertyBell]

On 2/20/2019 at 8:36 AM, 40/70 Benchmark said:

Let me preface this by saying that Raindance has had the best outlook thus far and mine has been bad. However given that this season's la nina like forcing regime has eaten the consensus alive, I'm not sure that I'd be extolling the double weighting of any el nino analog as a virtue of any outlook.

I'll leave it at that.

I think we pay too much emphasis on ENSO when there are many more factors that are more influential unless ENSO is strong and overwhelms the pattern.  We've had several la ninas with very snowy winters and several el ninos that lacked snowfall.

 

[CoastalWx]

21 hours ago, 40/70 Benchmark said:

Interesting...well, that explains alot. Tricky proposition forecasting sensible weather outcomes with blatant disregard for upper air patterns. I don't think I'll venture to try that.

If you don’t care about upper air pattern, it’s tough to take someone who is into long range forecasting, seriously.

[raindancewx]

The problem with using upper air patterns too heavily is everyone looks for canonical ENSO patterns to develop, instead of looking at what is actually happening on the ground. My interest is in getting the seasonal totals correct through math, data mining and statistics. The upper air patterns can be ideal for a season and you can still have weather completely different from what you expect. It's particularly difficult in the West, since the upper air patterns can be a match but the cold air or warm air may or may not get over the mountains. That's why long range forecasts do so much better in the East than in the West. 

Anyway, I came in to post that with the snow today, pending the official numbers Albuquerque is probably at or very near average seasonal snowfall totals. In El Nino years, that is strongly negatively correlated with heavy snowfall in the NE coastal plain. Best El Ninos for snow out here include years like 1972, 1997, 2006 which are lousy in the NE.

[LibertyBell]

6 hours ago, raindancewx said:

The problem with using upper air patterns too heavily is everyone looks for canonical ENSO patterns to develop, instead of looking at what is actually happening on the ground. My interest is in getting the seasonal totals correct through math, data mining and statistics. The upper air patterns can be ideal for a season and you can still have weather completely different from what you expect. It's particularly difficult in the West, since the upper air patterns can be a match but the cold air or warm air may or may not get over the mountains. That's why long range forecasts do so much better in the East than in the West. 

Anyway, I came in to post that with the snow today, pending the official numbers Albuquerque is probably at or very near average seasonal snowfall totals. In El Nino years, that is strongly negatively correlated with heavy snowfall in the NE coastal plain. Best El Ninos for snow out here include years like 1972, 1997, 2006 which are lousy in the NE.

Looks like Flagstaff has done extremely well, with that three foot plus snowstorm.  Gorgeous city with advanced lighting techniques to keep light pollution down and the high elevation means the air is clean too (very low asthma rates.)

https://en.wikipedia.org/wiki/Flagstaff,_Arizona#Dark_Sky_City

The amount of snow in the West has been immense and correlates well with the trough west ridge east pattern that's dominated this season.  Record warmth in the southeast this month too, that SE Ridge has been persistent since last summer when we doubled our previous 75+ dew point record here (42).

I was actually going to ask you about big snowfall years in Flagstaff vs the I-95 corridor but it looks like you already answered my question (Albuquerque is a good-enough proxy lol.)

This winter once again demonstrates why the connection between temperatures and snowfall isn't a strong one, we've had plenty of cold winters that had below average snowfall and plenty of mild winters that had at or above average snowfall.  It's all about the predominant pattern.  Even with above average precipitation, the dominant pattern has favored dry/cold alternating with wet/warm.

 

 

 

[MJO812]

On 2/21/2019 at 7:48 AM, NorEastermass128 said:

Now that futility is over, let’s shoot for normal!  

Not going to happen

Terrible winter for Philly NYC and Boston 

DC will end up near normal

[CoastalWx]

19 hours ago, raindancewx said:

The problem with using upper air patterns too heavily is everyone looks for canonical ENSO patterns to develop, instead of looking at what is actually happening on the ground. My interest is in getting the seasonal totals correct through math, data mining and statistics. The upper air patterns can be ideal for a season and you can still have weather completely different from what you expect. It's particularly difficult in the West, since the upper air patterns can be a match but the cold air or warm air may or may not get over the mountains. That's why long range forecasts do so much better in the East than in the West. 

Anyway, I came in to post that with the snow today, pending the official numbers Albuquerque is probably at or very near average seasonal snowfall totals. In El Nino years, that is strongly negatively correlated with heavy snowfall in the NE coastal plain. Best El Ninos for snow out here include years like 1972, 1997, 2006 which are lousy in the NE.

It's been good in the northeast outside of BOS-PHL though. And the H5 patterns are certainly needed with the help looking at all factors, not just hoping for canonical looks. If you are going to monitor every .05C change of region 3 and calculate the SOI to predict the pattern 4 months out, good luck.  The pattern of your 4 years there has looked nothing like what we have had. 

[raindancewx]

I agree, but we haven't as much snow as those years. 1972-73 finished with 37.4" here,  2006-07 was 27"+ and 1997-98 was higher too by now. When I went with low snow on the NE, I based it on the analogs I used for snow patterns - the heavy snow in the NE (Nov) and NW (Feb) for instance showed up in the analogs. My analogs actually did have Nino 3.4 as a 27.4C El Nino for Dec-Feb, which looks pretty close, it's actually much easier to figure stuff out that stuff. November 1953 is the only low solar El Nino with big snows in the coastal NE, and it turned warm in February in 1954.

And then went with this - 

So far...this is what we have.  I'd give myself an B for Dec-Jan nationally. My snow map looks pretty good to me so far, so I did something right. February is flipped for temps from what I expected but it is helping with the precip aspect. 

[40/70 Benchmark]

@raindancewx I still don't totally buy into your methodolgy, but you obviously did something right this season. Good job overall.

Look forward to your thoughts next year.

[TauntonBlizzard2013]

It’s an F.... this winter has blown.

[SnoSki14]

1 hour ago, TauntonBlizzard2013 said:

It’s an F.... this winter has blown.

Worst winter ever for me when you take into account slightly AN temps (N-J), AN precip, preseason forecasts, November tease, all snow to rain systems, and those atrocious weeklies. 

[raindancewx]

This is pretty rough, but this is essentially how I score my overall seasonal forecasts for highs - the red areas I consider big misses on my part. I should mention: When I say I have regressions that are pretty powerful - they are. But for the season. The regression stuff doesn't really work for individual months. I try to get the month with different methods. The red area in the North may be a bit deeper to the SW, I have to the look. Billings is the worst spot I've found on the high blend.

DJF High (F) Verification (12/1-2/22)	1953	1976	1986	1994	1994	2006	Mean	Actual	Error	Score
Atlanta	55.7	46.7	52.4	55.9	55.9	56.3	53.8	56.9	3.1	65.0
Albuquerque	50.4	47.7	46.7	53.4	53.4	45.6	49.5	47.6	-1.9	85.0
Amarillo	55.1	50.6	48.6	54.8	54.8	46.7	51.8	54.3	2.5	75.0
Billings	39.5	38.3	43.5	41.3	41.3	36.5	40.1	33.4	-6.7	35.0
Bismarck	28.7	22.4	34.2	24.2	24.2	27.5	26.9	22.4	-4.5	55.0
Boston	41.2	33.9	38.1	41.4	41.4	40.5	39.4	41.5	2.1	75.0
Des Moines	38.8	30.5	38.3	33.9	33.9	34.2	34.9	32.6	-2.3	75.0
Denver	50.7	49.2	45.8	49.1	49.1	37.4	46.9	44.2	-2.7	75.0
Detroit	37.3	26.8	34.8	36.5	36.5	35.4	34.6	36.1	1.6	85.0
Evansville	47.1	35.4	42.5	45.1	45.1	44.1	43.2	45.1	1.9	85.0
Elkins	43.7	33.7	40.5	43.0	43.0	41.3	40.9	43.8	2.9	75.0
El Paso	60.0	58.5	57.4	62.4	62.4	56.8	59.6	59.3	-0.3	100.0
Jacksonville	67.8	59.4	64.5	65.4	65.4	68.2	65.1	69.3	4.2	55.0
Lafayette	65.4	59.3	60.5	64.0	64.0	62.0	62.5	65.9	3.4	65.0
Philadelphia	45.3	34.3	41.5	45.1	45.1	44.3	42.6	44.1	1.5	85.0
Reno	49.8	50.7	48.7	49.2	49.2	48.0	49.3	46.3	-3.0	75.0
Salt Lake City	41.0	41.3	38.8	44.0	44.0	38.1	41.2	38.1	-3.1	65.0
San Diego	67.5	70.4	66.1	64.8	64.8	64.3	66.3	65.4	-0.9	95.0
Seattle	45.2	49.8	48.3	49.8	49.8	46.1	48.2	47.6	-0.6	95.0
St. Louis	47.3	34.8	42.2	42.5	42.5	42.5	42.0	42.8	0.8	95.0
Topeka	46.1	38.5	43.1	43.0	43.0	42.3	42.7	41.4	-1.3	85.0
Selected Cities Absolute Average Error									2.4	75.0

[Cold Miser]

Right now I am on a decent track to be lower than 11/12 with futility. 

[Hoth]

Yeah, bring on Morch. The toxic cocktail of high expectations going into the season, teases of beautiful patterns in the weeklies that kept failing, bad timing, Scooter Streaks effing up the December and January opportunities in favor of the MA, mixed bag events, way below average snow despite greater than average precip and favorable temp regimes, and the sheer stability of this garbage all winter, makes this one of the worst in my life. I was too young to remember the 80s and early 90s up until March '93, so I'll say it's the worst in recent memory. This one even beats '11-'12, which at least featured one exotic system and a historic heatwave. '15-'16 was a stinker too, but frankly I was still satiated by the glory of the previous season. It is what it is. I'm over it. We regroup and look forward to warmer weather, beaches, babes and barbecues.

[CoastalWx]

10 hours ago, raindancewx said:

I agree, but we haven't as much snow as those years. 1972-73 finished with 37.4" here,  2006-07 was 27"+ and 1997-98 was higher too by now. When I went with low snow on the NE, I based it on the analogs I used for snow patterns - the heavy snow in the NE (Nov) and NW (Feb) for instance showed up in the analogs. My analogs actually did have Nino 3.4 as a 27.4C El Nino for Dec-Feb, which looks pretty close, it's actually much easier to figure stuff out that stuff. November 1953 is the only low solar El Nino with big snows in the coastal NE, and it turned warm in February in 1954.

And then went with this - 

So far...this is what we have.  I'd give myself an B for Dec-Jan nationally. My snow map looks pretty good to me so far, so I did something right. February is flipped for temps from what I expected but it is helping with the precip aspect. 

Sometimes the Farmers Almanac can be right. 

[ORH_wxman]

I'd give the raindance snowmap a C/C-. Got the above average snows in central plains/eastern LakesN VT/southwest US correct....completely whiffed on the big snows in the southern mid-Atlantic, Maine, Sierras, and northern Rockies. Overpredicted snows in the southern plains.

The predicted snowmap looks like a +PNA El Niño with a STJ undercutting it. (Hence the below normal snows northern Rockies and above average southwest) And then no blocking which keeps everyone on the east coast below average. But reality didn't turn out that way at all. 

Snowfall is really hard to predict.