January 2017 Discussion & Observations

[gravitylover]

8 minutes ago, Allsnow said:

We are lucky to have Don postings here 

We've been lucky to have Don posting and as a friend of the boards for a lot of years.  Thanks Don 

[EasternLI]

Is it just me, or was that a fairly fantastic Euro weeklies run? Looks like after a relax, it goes Aleutian low and +PNA till the end.

[donsutherland1]

23 minutes ago, mimillman said:

Don-- do you have a similar PDF for NYC snowfalls 4+? I am trying to create similar analyses as you going forward and am compiling the data.

I had compiled a list of such storms in January. The list is below:

[mimillman]

7 minutes ago, donsutherland1 said:

I had compiled a list of such storms in January. The list is below:

Thanks. After looking at the data, I'm actually a bit concerned by the validity of how much we can read from the results. All indices appear to exhibit extreme auto-correlation. I haven't looked at the storm data as closely but I would assume it does as well. Still interesting information to have in terms of frequency distribution.

[donsutherland1]

5 minutes ago, mimillman said:

Thanks. After looking at the data, I'm actually a bit concerned by the validity of how much we can read from the results. All indices appear to exhibit extreme auto-correlation. I haven't looked at the storm data as closely but I would assume it does as well. Still interesting information to have in terms of frequency distribution.

One shouldn't take the indices too literally. They provide quick insight into the overall synoptic pattern, but the actual details e.g., ridge-trough placement, matter far more than the Index values. 

For example, Washington DC and New York City haven't seen many significant snowstorms (6" or more) since 1950 when the AO was positive and the PNA was negative (DCA needs blocking at least as much as NYC does given its somewhat warmer climate). Yet, this weekend will see a moisture-laden storm moving off the Southeast Coast so the risk that cities such as Richmond and Washington could see 6" or more will exist. The storm track, which will be a function of the synoptic details, will determine whether those cities (or even NYC) has a significant snowfall regardless of what the Index values are at the time. 

[mimillman]

3 hours ago, donsutherland1 said:

For those who are interested, here's a scatter diagram of NYC's 6" or greater January snowstorms:

 

For consistency, I thought I would also show the PNA vs. the NAO for some additional context.

 

[mimillman]

56 minutes ago, donsutherland1 said:

One shouldn't take the indices too literally. They provide quick insight into the overall synoptic pattern, but the actual details e.g., ridge-trough placement, matter far more than the Index values. 

For example, Washington DC and New York City haven't seen many significant snowstorms (6" or more) since 1950 when the AO was positive and the PNA was negative (DCA needs blocking at least as much as NYC does given its somewhat warmer climate). Yet, this weekend will see a moisture-laden storm moving off the Southeast Coast so the risk that cities such as Richmond and Washington could see 6" or more will exist. The storm track, which will be a function of the synoptic details, will determine whether those cities (or even NYC) has a significant snowfall regardless of what the Index values are at the time. 

I agree, I merely mean to point out that cluster analyses and scatter plots, while helpful, will tend to highly skew data exhibiting auto-regressive tendencies. That being said, I like to look at them myself

Edit: and sorry one last thing and I promise this inner-nerd outbreak will end. I want to show you exactly what I mean by this auto-correlation. Below, I show the NAO (as an example) ACF plot. The ACF plot tests auto-correlation of the series at varying lag periods. As you can see, there is statistically significant correlation out as far as Day 90 (3-month correlation), though the correlation significantly degrades after Day 20. 

I also show the results of the best ARIMA model. The model indicates the NAO time series is best described by an auto-regressive 5th degree moving average with 1st difference.

Series: nao 
ARIMA(0,1,5)                    

Coefficients:
          ma1      ma2      ma3      ma4      ma5
      -0.1395  -0.1970  -0.1789  -0.1440  -0.1235
s.e.   0.0065   0.0065   0.0072   0.0071   0.0069

sigma^2 estimated as 436677:  log likelihood=-192032.2
AIC=384076.3   AICc=384076.3   BIC=384124.9

 

[donsutherland1]

1 hour ago, mimillman said:

I agree, I merely mean to point out that cluster analyses and scatter plots, while helpful, will tend to highly skew data exhibiting auto-regressive tendencies. That being said, I like to look at them myself

They can. One should be aware of limitations when using the teleconnections or other indices e.g., don't look for too much precision. One can take a sample of cases where, for example, the PNA is +1 to +2 and then find differences in the overall synoptic pattern. Even similar 500 mb patterns can yield different results.

P.S. Very useful illustration concerning the NAO's autocorrelation. One finds autocorrelation in a lot of meteorological values e.g., temperatures.

[Morris]

Only 0.18" to show at the park after raining all day.

[weathermedic]

.35 at my station in Sheepshead Bay Brooklyn 

[psv88]

0.28

[Rtd208]

Picked up 0.42" of rain for the day.

[thunderbolt]

2 hours ago, donsutherland1 said:

They can. One should be aware of limitations when using the teleconnections or other indices e.g., don't look for too much precision. One can take a sample of cases where, for example, the PNA is +1 to +2 and then find differences in the overall synoptic pattern. Even similar 500 mb patterns can yield different results.

P.S. Very useful illustration concerning the NAO's autocorrelation. One finds autocorrelation in a lot of meteorological values e.g., temperatures.

do you look at the indices 3 days prior to the storm hitting weather increasing or decreasing

[donsutherland1]

5 hours ago, thunderbolt said:

do you look at the indices 3 days prior to the storm hitting weather increasing or decreasing

I watch all aspects of the pattern, including the indices. Changes could suggest a sharpening or broadening trough, among other things, which would have implications for a possible storm.

[bluewave]

Once we lose the transient cold shot January 5-10th, the weeklies run the mild weather right into later January.

The -PNA trough retrogrades back to Alaska and flips the EPO positive. It will be interesting to see if the

weeklies are correct about the shift to a +PNA around January 25-27. This is about a week later than the

last weekly run was showing. Very strong agreement on this pattern evolution through mid-January on the

GEFS and CMC ensembles also. 

[Allsnow]

9 minutes ago, bluewave said:

Once we lose the transient cold shot January 5-10th, the weeklies run the mild weather right into later January.

The -PNA trough retrogrades back to Alaska and flips the EPO positive. It will be interesting to see if the

weeklies are correct about the shift to a +PNA around January 25-27. This is about a week later than the

last weekly run was showing. Very strong agreement on this pattern evolution through mid-January on the

GEFS and CMC ensembles also. 

Yeah weeklies looked great for the end of January and The start of February. But after this week its break out the golf clubs in mid January 

[bluewave]

9 minutes ago, Allsnow said:

Yeah weeklies looked great for the end of January and The start of February. But after this week its break out the golf clubs in mid January 

The weeklies after January 25th looked like that strong -EPO/+PNA Feb CFS run that has gotten posted a few times. I wonder if the models

are onto something or it just gets pushed further back the closer we get? I know the models can often rush these type of pattern changes.

But the pattern before that looks as La Nina as you can get with the big vortex pulling back to Alaska.

 

[Googlymoogly]

How mild are we gonna get?

 

[PB GFI]

1 hour ago, bluewave said:

The weeklies after January 25th looked like that strong -EPO/+PNA Feb CFS run that has gotten posted a few times. I wonder if the models

are onto something or it just gets pushed further back the closer we get? I know the models can often rush these type of pattern changes.

But the pattern before that looks as La Nina as you can get with the big vortex pulling back to Alaska.

 

 

I posted the Dec 22  2 M weeklies ending the 11th away .

Check out the new day 4 thru 11 ending the 14th 

 

The weeklies busted so bad in just 2 weeks . They are useless.

 

That said I am fully on board with the 20 thru 30 being AN and if it starts on the 15th so be it .

 

We lose the EPO and WPO so its game over come mid month .

 

But there are 2 chances for snow over the next 6 days and thats going to make or break Jan for snow .

 

 

[donsutherland1]

At 8 am, the Central Park temperature was 39°. It seems to have "flat-lined" there, as it has now been 39° for 15 consecutive hours.

[Allsnow]

44 minutes ago, Googlymoogly said:

How mild are we gonna get?

 

If the weeklies are correct we torch 

[Brian5671]

1 hour ago, donsutherland1 said:

At 8 am, the Central Park temperature was 39°. It seems to have "flat-lined" there, as it has now been 39° for 15 consecutive hours.

same story at my residence, we've been 37-38 degrees last day or so.

[WEATHERBUFF]

1 hour ago, Allsnow said:

If the weeklies are correct we torch 

The weeklies have been inconsistent.

[WEATHERBUFF]

It seems like 2 top analogs in play for this winter 98-99 and 58-59. I was originally thinking 83-84 but the way things are shaping up I think the first 2 do have some legs here.

[pazzo83]

2 hours ago, donsutherland1 said:

At 8 am, the Central Park temperature was 39°. It seems to have "flat-lined" there, as it has now been 39° for 15 consecutive hours.

We have a pretty remarkable stretch ongoing right now - outside of an hour or so early Saturday morning, we've been at or above freezing (and really only at freezing for a few hrs Sat morning) since the morning of Wed, Dec 21st.

[Stormlover74]

14 minutes ago, pazzo83 said:

We have a pretty remarkable stretch ongoing right now - outside of an hour or so early Saturday morning, we've been at or above freezing (and really only at freezing for a few hrs Sat morning) since the morning of Wed, Dec 21st.

Last year we were at or above freezing from March 30, 2015 until January 4, 2016

[Brian5671]

1 hour ago, WEATHERBUFF said:

It seems like 2 top analogs in play for this winter 98-99 and 58-59. I was originally thinking 83-84 but the way things are shaping up I think the first 2 do have some legs here.

98-99 was horrid here.  Not sure how 58-59 worked out.

[donsutherland1]

3 hours ago, Allsnow said:

If the weeklies are correct we torch 

The weekly guidance hasn't been great. Having said that, there is a strong signal for warmer than normal readings in the East around mid-month on the ensembles, objective analogs, and even based on the forecast teleconnection indices (the AO- notwithstanding). Some maps:

 

[PB GFI]

22 minutes ago, donsutherland1 said:

The weekly guidance hasn't been great. Having said that, there is a strong signal for warmer than normal readings in the East around mid-month on the ensembles, objective analogs, and even based on the forecast teleconnection indices (the AO- notwithstanding). Some maps:

 

 

The 20th - 30th are AN , TBF it prob starts around just after the 15th 

No question / we will turn mild  once again . 

 

 

 

[wthrmn654]

10 minutes ago, PB GFI said:

 

The 20th - 30th are AN , TBF it prob starts around just after the 15th 

No question / we will turn mild  once again . 

 

 

 

Shall we call that the January thaw,our whatever they usually call it?