Texas/New Mexico/Louisiana/Mexico Obs And Discussion Thread Part 8

[bubba hotep]

47 minutes ago, aggiegeog said:

Looks like today featured sub freezing highs for a large chunk of the state. Highest here was right around 30 with wind child in the 10s for much of the day. Can we reach the mid teens tonight with lighter winds and less clouds? This morning was 21 here.

I think the high IMBY was around 26 based on the old fashion thermometer that came with my rain gauge. My NWS point forecast is 12 but the last couple of cold nights have busted low, single digits maybe? 

[DoctorMu]

12 hours ago, bubba hotep said:

Just a slight change from 00z to 06z...

 

 

Just...a bit outside!

[Quixotic1]

Little anecdote from this morning: haven't been able to cut the yard because of travel, work and school but the grass really hasn't been growing.  This morning I went out for firewood and there's a big dandelion about 2 feet tall.  I thunked it and it broke like glass. 

[Quixotic1]

Little anecdote from this morning: haven't been able to cut the yard because of travel, work and school but the grass really hasn't been growing.  This morning I went out for firewood and there's a big dandelion about 2 feet tall.  I thunked it and it broke like glass. 

[1900hurricane]

Wow, 00Z ECMWF actually looks like it it could be a Christmas Day chasing miracle across the plains!

[bubba hotep]

13 hours ago, 1900hurricane said:

Wow, 00Z ECMWF actually looks like it it could be a Christmas Day chasing miracle across the plains!

Still a week for things to play out but it certainly brings up memories of last year and the Garland, TX tornado on the 26th. It crossed I30 about 2 or 3 minutes before we got off and onto George Bush. We saw a bunch of power flashes and there was debris all over the road. For anyone that knows that area, there were huge pieces of debris thrown up on George Bush where it exits off I30 and heads out over the lake. 

[raindancewx]

So it turns out that having intense warmth in the upper Midwest/Montana in December is a good sign for big snows in the Southwest in March...sure enough it's super cold up there this year. Not good.

[bubba hotep]

Bleh... Looks like a warmer period until we get another -EPO in early January. How deep and long are the questions. However, it does look like we will get some rain until then. Cold dry and warm wet is no way to run a winter! 

[raindancewx]

1983 continues to look really strong as an analog in my area - we have a real shot at a white Christmas here. Since 1931 as far as I can tell the city has only had four measurable snows on Christmas, 1939, 1962, 1974, and 1983. 

1974 is by far the least crappy of all La Ninas out here for snow - 16.5" snow in ABQ.

[bubba hotep]

Uh oh, major cold coming?

[bubba hotep]

11 minutes ago, bubba hotep said:

Uh oh, major cold coming?

Looks kind of familiar 

[Quixotic1]

nice to see some west nao blocking......

[raindancewx]

NOAA likes the middle of January for wetness in much of the West.

Here is hoping it verifies, snow pack here is only ~77% of normal even with the recent storms.

[aggiegeog]

Early January still appears to be the next target for winter weather in this region. 

[bubba hotep]

On 12/22/2016 at 7:54 AM, bubba hotep said:

Looks kind of familiar 

The 12z GFS today is an even more classic look, too bad it is way out in fantasy land.

[raindancewx]

I was playing around with projecting seasonal snow totals for Albuquerque the other day, and I came up with six variables that account for ~78% (r squared = 0.78) of the variation in snow in a given Oct-May season here. They are hard to figure out, but have been asking people on Twitter, and they've done a decent job so far - quite close on first snow date. Essentially:

a) ONI in DJF (I used ONI estimates from 1931-32 to 1949-50 also)

Precipitation in DJF

c) Mean High in DJF

d) Precipitation in August + Precipitation in October (as one variable - combined)

e) First measurable snow as a date (Oct 1 = 1, Oct 2 = 2, etc, measurable means >=0.1" snow in 24 hrs ending midnight)

f) Snowy Day Frequency (how many days had >=0.1" snow from Oct-May, day is 24 hrs ending midnight)

ONI has some effect on most of the other variables, particularly b, d, and f. Weak effect on c too. F is the strongest correlation to snow overall, but adding the other variables helps a lot, because you can have 10 days with each day averaging an inch...or two inches...or half an inch. Variable B/D have some control over frequency. 

Generally:

A --> Strong relationship to total snow, and if the top month in the season is very snowy, i.e. El Ninos usually have one big snow month here, which leads to a big seasonal total.

B --> Reinforces/contradicts the frequency variable, and also adjusts intensity of that variable, i.e 15 snows in a winter with 0.70" precipitation won't be that great.

C --> Some influence on snow intensity on the days it snows, via snow ratios (15:1 more likely if cold?), and some influence on Spring snow likelihood.

D --> Lots of rain, lots of snow in Jan-Apr, especially March, and overall. March snow virtual lock if rainy. Little rain, snow shuts down fast after December. In between...can go either way.

E --> Literally an indicator of whether it snowed in November or not. Snow in November usually indicates a snowier than average Oct-May period. No snow by mid-Dec usually indicates average to below average snow. Roughly speaking, snow by Nov 10 - OK, first snow Nov 11 - 30 - Ideal, first snow Dec 1-14 - OK, first snow after Dec 14 - bad.

F --> More is better, but how much better depends on the other factors. With more snows, odds of a big snow from one system increase too.

An ideal snow season, according to the model, that is possible based on conditions that have been observed in the past 85 years for each of the six variables, would produce 47 inches of snow from October to May - would need a big El Nino, with early snow, lots of rain in August & October, lots of days with snow, a cold winter, and a wet winter.

The long & the short of this is some of the factors look **better** than last year despite the La Nina. We had an earlier first snow at the airport, and more rain in August & October. Both support more snow in the shoulder months (Nov/Mar) - it did snow a tiny bit in Nov. Suspect this is a colder winter than last year too which helps, as it probably won't torch in February for a third year in a row, and January looks average to cold, with December mild/warm.

This is what the hindcast v. observed discrepancies look like overall. Average error is +/-2.2", 87% shot at being within 4 inches in 85 years.

[Quixotic1]

3 hours ago, raindancewx said:

I was playing around with projecting seasonal snow totals for Albuquerque the other day, and I came up with six variables that account for ~78% (r squared = 0.78) of the variation in snow in a given Oct-May season here. They are hard to figure out, but have been asking people on Twitter, and they've done a decent job so far - quite close on first snow date. Essentially:

a) ONI in DJF (I used ONI estimates from 1931-32 to 1949-50 also)

Precipitation in DJF

c) Mean High in DJF

d) Precipitation in August + Precipitation in October (as one variable - combined)

e) First measurable snow as a date (Oct 1 = 1, Oct 2 = 2, etc, measurable means >=0.1" snow in 24 hrs ending midnight)

f) Snowy Day Frequency (how many days had >=0.1" snow from Oct-May, day is 24 hrs ending midnight)

ONI has some effect on most of the other variables, particularly b, d, and f. Weak effect on c too. F is the strongest correlation to snow overall, but adding the other variables helps a lot, because you can have 10 days with each day averaging an inch...or two inches...or half an inch. Variable B/D have some control over frequency. 

Generally:

A --> Strong relationship to total snow, and if the top month in the season is very snowy, i.e. El Ninos usually have one big snow month here, which leads to a big seasonal total.

B --> Reinforces/contradicts the frequency variable, and also adjusts intensity of that variable, i.e 15 snows in a winter with 0.70" precipitation won't be that great.

C --> Some influence on snow intensity on the days it snows, via snow ratios (15:1 more likely if cold?), and some influence on Spring snow likelihood.

D --> Lots of rain, lots of snow in Jan-Apr, especially March, and overall. March snow virtual lock if rainy. Little rain, snow shuts down fast after December. In between...can go either way.

E --> Literally an indicator of whether it snowed in November or not. Snow in November usually indicates a snowier than average Oct-May period. No snow by mid-Dec usually indicates average to below average snow. Roughly speaking, snow by Nov 10 - OK, first snow Nov 11 - 30 - Ideal, first snow Dec 1-14 - OK, first snow after Dec 14 - bad.

F --> More is better, but how much better depends on the other factors. With more snows, odds of a big snow from one system increase too.

An ideal snow season, according to the model, that is possible based on conditions that have been observed in the past 85 years for each of the six variables, would produce 47 inches of snow from October to May - would need a big El Nino, with early snow, lots of rain in August & October, lots of days with snow, a cold winter, and a wet winter.

The long & the short of this is some of the factors look **better** than last year despite the La Nina. We had an earlier first snow at the airport, and more rain in August & October. Both support more snow in the shoulder months (Nov/Mar) - it did snow a tiny bit in Nov. Suspect this is a colder winter than last year too which helps, as it probably won't torch in February for a third year in a row, and January looks average to cold, with December mild/warm.

This is what the hindcast v. observed discrepancies look like overall. Average error is +/-2.2", 87% shot at being within 4 inches in 85 years.

Nice work.  I would recommend checking for overfit of your model.  I've got some background for you if you're interested.  I had a model at work that had a ridiculous r squared but was overfit.  I did some digging around and found out the original data, unbeknownst to me, had been in triplicate.  Creating overfit.  Anywho, I'd be happy to take a look at it for you if you're interested.  I think Judah Cohen and the Italian guys fell victim to it.  

Cheers.  

[DFWWeather]

21 hours ago, bubba hotep said:

The 12z GFS today is an even more classic look, too bad it is way out in fantasy land.

0z today is as wickedly cold. It has nearly -15°C air sitting on top of DFW with surface temps in the single digits with highs only in the upper teens to around 20 for the first week of January. La La Land....ECMWF has nothing at all. However, EPO, WPO are projected to go strongly negative. The PNA looks to also be in the negative territory (which isn't a great signal for us down south). It does appear that the pattern that brought us cold in early December may try to come again with ridging into Alaska. This may be a colder version of that pattern though.

[Severe Wx]

Still looking good for thunderstorms in north texas today?

[bubba hotep]

Models seem to be coming into agreement on a bit more sustained cold blast b/w Jan 6 - 10th. Obviously, details still need to be worked out but a -EPO driven cold shot looks to be in the cards. The models have been pretty good at sniffing out these -EPO patterns in the longer range. 

[bubba hotep]

I'll never forget this, wasn't out chasing, but ended up way too close for comfort.

[bubba hotep]

12z GFS with a bit of frozen for DFW next week:

[ams30721us]

4 hours ago, bubba hotep said:

12z GFS with a bit of frozen for DFW next week:

 

Looks like both the 12z GFS & 12Z Euro Suite are on board for getting some wintry weather down to at least the I-20 corridor in Texas by the end of next week behind the next arctic front.

[ams30721us]

Oh...and let's be fair and include the CMC too at least in some respect.

[bubba hotep]

Still there at 18z

[aggiegeog]

Next week she is looking good for winter lovers. Arctic air with a fairly zonal flow allowing for moisture to ride over the surface cold. Sadly it looks more icy than snowy at this point. 

[bubba hotep]

2 hours ago, aggiegeog said:

Next week she is looking good for winter lovers. Arctic air with a fairly zonal flow allowing for moisture to ride over the surface cold. Sadly it looks more icy than snowy at this point. 

The 00z Euro package was one of the snowiest yet this winter for the northern half of Texas. However, as you noted, that is probably mostly freezing rain given the setup. 

[bubba hotep]

Both the Euro EPS and GEFS show this being a longer cold outbreak than the last one, lasting 5 days to a week before a real warm up. They also show our source region staying below normal in the longer range and that might be an indication that another dump is coming after any warm up. 

[jhamps10]

12z Gfs takes the storm away until the weekend of the 7th-8th.. Shows what would probably be a big freezing rain or sleet event in dfw with temps in the mid 20s.. Takes away the overrunning. Still wouldn't  shock me to see it get icy next week, and not just the temps

[bubba hotep]

Another variation of a system that will most likely never come to fruition...