January Storm Potentials

[HM]

Interesting post as always HM. Definitely going to have much more of a ridge out west to work with for the Christmas threat. However, the pesky trough crashing onto the west coast, and the ridge axis shifting east over Montana instead of Boise Idaho around storm day is concerning that this will all come together too far east once again. Or is this where the shorter wavelengths that HM mentioned can come into play and still work out for an east coast storm?

Great point about the ridge out west; and while it may be a little too east for my liking, it certainly is more amplified than the current situation. You also won't have this massive crushing summer-like 570dm ridge over Baffin Island either with a PV squeezing underneath. However, we are still dealing with PV dynamics with this storm on Christmas and I hate that. They fail about 75% of the time. I like the classic setup where a 50-50 low is present and there is nothing else standing between that and the incoming s/w to become a monster storm. I could see the Christmas system not being the big one and something after it becomes the big one; although, I would urge caution forecasting anything big in a strong La Niña. It is just rare to happen...it can...but rare.

[NJwinter23]

Great point about the ridge out west; and while it may be a little too east for my liking, it certainly is more amplified than the current situation. You also won't have this massive crushing summer-like 570dm ridge over Baffin Island either with a PV squeezing underneath. However, we are still dealing with PV dynamics with this storm on Christmas and I hate that. They fail about 75% of the time. I like the classic setup where a 50-50 low is present and there is nothing else standing between that and the incoming s/w to become a monster storm. I could see the Christmas system not being the big one and something after it becomes the big one; although, I would urge caution forecasting anything big in a strong La Niña. It is just rare to happen...it can...but rare.

http://www.meteo.psu.edu/~gadomski/NARR/1996/us0106.php

but I digress.. Great points and I'd assume a big reason why these long range threats involving PV dynamics fail most of the time is because the ingredients all come together perfectly on the models 5-6-7 days out when we first recognize the threat, but in reality that interaction is usually going to be off at that range. And of course this winter strong la nina climo is the 800lb gorilla in the room against these monster gfs spinups.

I do think because of the amplification ongoing that we will at least see a moderate-significant winter storm progressing east from the Midwest in time for Xmas day. Whether that produces in the mid-Atl, northeast or both we shall see. I am definitely not falling for the gfs blizzard.

Though we don't always need the ridge centered in the perfect spot for a KU:

http://www.meteo.psu.edu/~gadomski/NARR/2003/us1205.php

[donsutherland1]

Yes, that is true, it has happened before (though I would say the blocking this past winter and this winter so far is more impressive overall). But all of those winters were weak/moderate El Ninos.

I agree. As you noted earlier, something larger is probably contributing. My guess concerns a longer-term AO cycle. But there are a range of other possibilities.

[JBG]

Thanks HM for giving us some hope for snow. So far in NEPA this has been a waste of cold air.I think you said a while back that week 3 in January could be wintry and I was just wondering if that is still valid. I always enjoy reading your posts.

When there's life, there's hope. But for a snowy winter, maybe not.

[weathermanchild]

Great point about the ridge out west; and while it may be a little too east for my liking, it certainly is more amplified than the current situation. You also won't have this massive crushing summer-like 570dm ridge over Baffin Island either with a PV squeezing underneath. However, we are still dealing with PV dynamics with this storm on Christmas and I hate that. They fail about 75% of the time. I like the classic setup where a 50-50 low is present and there is nothing else standing between that and the incoming s/w to become a monster storm. I could see the Christmas system not being the big one and something after it becomes the big one; although, I would urge caution forecasting anything big in a strong La Niña. It is just rare to happen...it can...but rare.

Um just thought this might be relevant to the Christmas discussion... To me this essentially shows that the ensemble members want to strengthen the coastal low and change the anomaly structure across the North Atlantic pretty substantially. Last time I saw a rapid increase in this index like this was mid March of this year when the Mid Atlantic got creamed with a very strong nor easter.

good stuff man, keep it goin.

[HM]

A quick update before I head out:

The MJO pulse described in the original post is turning out to be quite impressive, initiating the Atlantic invest and Pac storm. The upward motion has been completely weakened across the Indonesian sector based on the latest OLR/CHI anomalies. It has also become quite apparent that the 12/25-26 storm is legit and likely the Heather A. signal. The slower progression of the s/w from the Southwest and monster blocking has completely allowed for the potential of a monster coastal storm. I look forward to tracking this one with all of you.

I'll be back for another evening of 00z runs.

[Jebman]

A quick update before I head out:

The MJO pulse described in the original post is turning out to be quite impressive, initiating the Atlantic invest and Pac storm. The upward motion has been completely weakened across the Indonesian sector based on the latest OLR/CHI anomalies. It has also become quite apparent that the 12/25-26 storm is legit and likely the Heather A. signal. The slower progression of the s/w from the Southwest and monster blocking has completely allowed for the potential of a monster coastal storm. I look forward to track this one with all of you.

I'll be back for another evening of 00z runs.

Whoever gets destroyed by this Xmas storm will truly be the shizzle. Ready your shovels and snowblowers folks! You heard HM!!! THE CHRISTMAS STORM IS THE ARCHAMBAULT STORM.

Take plenty of pictures! We are about to see a ton of radars and model outputs that will be legendary archive material! Get those hard drives ready! Never forget what happened to Eastern! Never forget how losing all that great storm info and commentary took fifty years off all our lives! WE ARE GONNA SPARE NO EFFORT TO ETERNALLY MEMORIALIZE THIS STORM!!

[NJwinter23]

A quick update before I head out:

The MJO pulse described in the original post is turning out to be quite impressive, initiating the Atlantic invest and Pac storm. The upward motion has been completely weakened across the Indonesian sector based on the latest OLR/CHI anomalies. It has also become quite apparent that the 12/25-26 storm is legit and likely the Heather A. signal. The slower progression of the s/w from the Southwest and monster blocking has completely allowed for the potential of a monster coastal storm. I look forward to tracking this one with all of you.

I'll be back for another evening of 00z runs.

I'll be waking up for work at 2 again to catch the end of another round. Hopefully the 3rd night in a row of a 00z euro blitzing. What an awesome signal this is growing into though. Check out that pineapple express into Socal and the southern Rockies right now. Anyone who is dismissing the major possibilities of this storm strictly because of mod/strong Nina climo need to get their head out of their arses and look what's going on with the pattern right now.

Now if you want to be skeptical because it's 5 days away still, then I can understand. I'm trying to remain reserved myself, but this looks really good right now.

[Rainshadow]

I'll be waking up for work at 2 again to catch the end of another round. Hopefully the 3rd night in a row of a 00z euro blitzing. What an awesome signal this is growing into though. Check out that pineapple express into Socal and the southern Rockies right now. Anyone who is dismissing the major possibilities of this storm strictly because of mod/strong Nina climo need to get their head out of their arses and look what's going on with the pattern right now.

Now if you want to be skeptical because it's 5 days away still, then I can understand. I'm trying to remain reserved myself, but this looks really good right now.

Well I'm posting this so HM can flame me (how else can I learn?), the gwo's latest turn is getting it closer to a la nina mode than it has been since October. The current quadrant that its in since 1958 there has not been a snowfall event in PHL of >10". I can't figure its movement completely and could it make the turn and start rising by the 26th and reach climatological occurrences, I can't say it won't. I don't know enough. But when it took this recent turn it gave me an oh-oh today. But in reality I'm just starting to scratch the surface on that and my skepticism is because its still 5 days away (chances are tomorrow it still will be 5 days away) and I'm always skeptical of totally phased solution in la nina winters beyond 96 hours. By the way the CIPS has had Dec 2000 as its favorite analog for a couple of days now, which really wasn't that bad of an event, especially toward NYC. Merry Christmas!

[SP]

Well I'm posting this so HM can flame me (how else can I learn?), the gwo's latest turn is getting it closer to a la nina mode than it has been since October. The current quadrant that its in since 1958 there has not been a snowfall event in PHL of >10". I can't figure its movement completely and could it make the turn and start rising by the 26th and reach climatological occurrences, I can't say it won't. I don't know enough. But when it took this recent turn it gave me an oh-oh today. But in reality I'm just starting to scratch the surface on that and my skepticism is because its still 5 days away (chances are tomorrow it still will be 5 days away) and I'm always skeptical of totally phased solution in la nina winters beyond 96 hours. By the way the CIPS has had Dec 2000 as its favorite analog for a couple of days now, which really wasn't that bad of an event, especially toward NYC. Merry Christmas!

Pulling off the moderate event or better would be like being down 21 points in the 4th quarter with less than 8 minutes left in the game, scoring 3 touchdowns in about 5 minutes and then running back a punt for the winning touchdown with no times left of the clock.

It only happens every 50 years or so.....

Merry Merry!

[Amped]

Well I'm posting this so HM can flame me (how else can I learn?), the gwo's latest turn is getting it closer to a la nina mode than it has been since October. The current quadrant that its in since 1958 there has not been a snowfall event in PHL of >10". I can't figure its movement completely and could it make the turn and start rising by the 26th and reach climatological occurrences, I can't say it won't. I don't know enough. But when it took this recent turn it gave me an oh-oh today. But in reality I'm just starting to scratch the surface on that and my skepticism is because its still 5 days away (chances are tomorrow it still will be 5 days away) and I'm always skeptical of totally phased solution in la nina winters beyond 96 hours. By the way the CIPS has had Dec 2000 as its favorite analog for a couple of days now, which really wasn't that bad of an event, especially toward NYC. Merry Christmas!

Yes we are in the low AAM phase, but it isn't that strong . Maybe only a moderate storm

[Rainshadow]

Yes we are in the low AAM phase, but it isn't that strong . Maybe only a moderate storm

I'm not greedy, I'd sign off on a moderate strong today.

[donsutherland1]

I'm not greedy, I'd sign off on a moderate strong today.

If I recall correctly, during La Niña events with an ENSO Region 3.4 anomaly of -1 or below in December, Phases 5-8 are ugly for PHL, while Phases 2-4 are best. Phase 1 has also had some snowfalls.

[HM]

Tony, I would never flame you. You bring up a valid point but fortunately (or unfortunately for some of us who don't want it) the damage has been done. The huge 50-50 low and western ridge response will be around long enough to allow for amplification and potential phasing. In the end, this comes down to the magic of phasing. I am more worried about the E PAC / west Ridge breaking down that ruins phasing over the 50-50 low crushing everything. Notice that the pattern, after the storm, grows more zonal. The GWO is certainly signaling the end; but once you got the blocks in place, it doesn't matter. Something that may trump a strong La Niña response would be an equally impressive retrograding -NAO block, which we have.

Going to be an interesting battle.

[HM]

If I recall correctly, during La Niña events with an ENSO Region 3.4 anomaly of -1 or below in December, Phases 5-8 are ugly for PHL, while Phases 2-4 are best. Phase 1 has also had some snowfalls.

I'd be interested to see this data but I don't think it matters much anyway. Sometimes there is a considerable lag in the effects of global AAM / tendency. Also, localized anomalies that propagate poleward from the equator may be more important than the overall global signal.

[Rainshadow]

I'd be interested to see this data but I don't think it matters much anyway. Sometimes there is a considerable lag in the effects of global AAM / tendency. Also, localized anomalies that propagate poleward from the equator may be more important than the overall global signal.

HM,

Thanks. I still worry though. On the Euro its still nearly 144 hours away (again). I'm not going to guess what phase the gwo is going to be come December 26th-27th (I'd be wrong anyway), if it so happens to still be in the "southern hemisphere" with an aam < -1 (I left the chart at work) and Philly gets 10" or greater, it would be a first since 1958 proving there are exceptions to every rule. Before it made this left turn it was heading toward the phases it was in during 1/96 & 1/2000.

[Rainshadow]

If I recall correctly, during La Niña events with an ENSO Region 3.4 anomaly of -1 or below in December, Phases 5-8 are ugly for PHL, while Phases 2-4 are best. Phase 1 has also had some snowfalls.

Don,

For next winter, I was going to look at the 6-9.9" events and add it to our local chart.

[Srain]

The pattern is looking good HM. Heck, we've got something in the forecast in our part of the world that we have not seen since November. Rain!

[Ender]

Don,

For next winter, I was going to look at the 6-9.9" events and add it to our local chart.

I think I've compiled this data for you, and everyone else...

At the bottom of this post there are two links to Excel workbooks (.xlsx type). The first Excel workbook, Discrete Snowfall Events, has all the >.1" snowfalls for each of RIC, IAD, DCA, BWI, PHL, NYC and BOS along with the AAM data as show here. Each event is listed by date along with the daily high/low, total liquid precip, snowfall, snow cover at the observation time, and finally the five AAM data points from the previously mentioned site.

The second workbook, Grouped Snowfalls, is the result of running the "discrete events" through a script that aggregates them into "storms". This is an attempt to show consecutive, multi-day accumulations as a single storm. I used a threshold of four days as the constraint for "longest storm duration". The first three columns reference the first date of measurable snow, the "peak date" (IOW, day with the highest accumulation in the set of days), and the last date of measurable snow in the series. The total precip is the sum of all the liquid measurements while the total snow is, obviously, the sum of the snow measurements. The AAM values should correspond to the value found on the "PEAK DATE".

Some caveats:

• I've only spot checked the values for each of the locations. There could be mistakes. Just let me know
  
• The beginning of time is 1/1/1900, so Excel likes to put observations from the 1800s at the bottom of an "oldest -> newest" sort.
  
• A lot of the pre 1960s snow cover measurements are pretty screwy. Either that or the script I used to originally pull those values is screwy. I haven't bothered to fix those errors and I probably won't. No, NYC didn't really have a 124" depth measurement.
  
• There are real gaps in some of the data. Those are shown as blanks and/or "M".
  

Discrete Snow Events

Grouped Snowfalls

[donsutherland1]

Outstanding work, Ender.

You have created an invaluable resource.

[winterymix]

I notice some of the dates show a high of 57 or 58, or in one case, 60, and I just have to

sit and meditate on that sort of day.

[JBG]

Some caveats:

• I've only spot checked the values for each of the locations. There could be mistakes. Just let me know
  

Great work.

On mistake that jumped out was that the February 4, 1995 Kocin storm didn't show up on the NYC stats. I'm sure there are others but that is a valuable resource. I'm sure Uncle WX1 would approve.

[Ender]

I notice some of the dates show a high of 57 or 58, or in one case, 60, and I just have to

sit and meditate on that sort of day.

Some times NCDC has garbage in their data, especially prior to the 1940s and especially with snow cover and low temps. Sometimes, however, there are column offsets that I didn't expect, or unexpected alphanumerics, that throw off my meager set of checks and data just gets dumped. I'm also using tons of nested hashes and hashes of hashes, and dynamic hashes, which aren't really supported, because I like to pretend I'm clever. Unfortunately that means I occasionally forget to reinitialize something and that can create issues.

Great work.

On mistake that jumped out was that the February 4, 1995 Kocin storm didn't show up on the NYC stats. I'm sure there are others but that is a valuable resource. I'm sure Uncle WX1 would approve.

I see that it made it into all the stations except NYC...for some reason I've got a bunch of "M"s in my NYC data from 2/2/95 - 2/28/95. The NYC data was a little goofy in a number of ways, I'll just go through it and spot check it more thoroughly. The NYC source data was one of the more messy sets, I'll try to figure out why.

I'd rather not derail HM's thread. If people would just PM me errors I'll take a look at them. The more errors I have the quicker I'll find the pattern, and the quicker I'll figure out which of my conditionals was stupid.

EDIT, it's not my stupidity (this time). The data I pulled from NCDC has a bunch of "-99999" for that month. I'll pull on of their other data sets, or I'll just reconstruct the daily data from the hourly OBS. Not sure I'm going to do that today, I haven't bought any gifts yet.

[HM]

Well, the MJO pulse is rounding the 60W to 60E belt, aiding in this potential major snowstorm for many locations. The situation is very similar to early November, as described in the first two posts. Notice that the GWO/AAM etc were all doing this then and produced the autumn version of what's ahead for us this weekend (you have to admit there is a strong similarity between the maps in post 2 and what's ahead).

A few pointers:

1. Just because you're not in the heaviest QPF, doesn't mean you will not see heavy snow. Systems like this always have a sick band on the NW edge near the dry air that produces more snow than indicated, especially when this thing begins to occlude/mature. Then if the wind isn't too disturbing, there is always the ratio factor.

2. I don't think we will see a monster gravity wave that cuts off snow like we did a couple years ago but we will see internal gravity waves with gusty winds and heavy snow, should this thing behave like the 00z GFS.

3. Snowfall rates will likely exceed 2"/hr in some of the CSI bands at the core of the max WAA/PVA.

Merry Christmas.

[Rainshadow]

Well, the MJO pulse is rounding the 60W to 60E belt, aiding in this potential major snowstorm for many locations. The situation is very similar to early November, as described in the first two posts. Notice that the GWO/AAM etc were all doing this then and produced the autumn version of what's ahead for us this weekend (you have to admit there is a strong similarity between the maps in post 2 and what's ahead).

A few pointers:

1. Just because you're not in the heaviest QPF, doesn't mean you will not see heavy snow. Systems like this always have a sick band on the NW edge near the dry air that produces more snow than indicated, especially when this thing begins to occlude/mature. Then if the wind isn't too disturbing, there is always the ratio factor.

2. I don't think we will see a monster gravity wave that cuts off snow like we did a couple years ago but we will see internal gravity waves with gusty winds and heavy snow, should this thing behave like the 00z GFS.

3. Snowfall rates will likely exceed 2"/hr in some of the CSI bands at the core of the max WAA/PVA.

Merry Christmas.

HM,

Pardon my gwo expressions, after all was said and done the gwo was shot out of a cannon and left the sw quadrant, its probably near the southern edge of where PHL has had 10" events in the past, I didn't get a chance to see much modeling the last three days (just the 12z run on Christmas Eve). Heather Archambault (I can think of some others too) must be a very happy person this morning.

[okie333]

Some times NCDC has garbage in their data, especially prior to the 1940s and especially with snow cover and low temps. Sometimes, however, there are column offsets that I didn't expect, or unexpected alphanumerics, that throw off my meager set of checks and data just gets dumped. I'm also using tons of nested hashes and hashes of hashes, and dynamic hashes, which aren't really supported, because I like to pretend I'm clever. Unfortunately that means I occasionally forget to reinitialize something and that can create issues.

I see that it made it into all the stations except NYC...for some reason I've got a bunch of "M"s in my NYC data from 2/2/95 - 2/28/95. The NYC data was a little goofy in a number of ways, I'll just go through it and spot check it more thoroughly. The NYC source data was one of the more messy sets, I'll try to figure out why.

I'd rather not derail HM's thread. If people would just PM me errors I'll take a look at them. The more errors I have the quicker I'll find the pattern, and the quicker I'll figure out which of my conditionals was stupid.

EDIT, it's not my stupidity (this time). The data I pulled from NCDC has a bunch of "-99999" for that month. I'll pull on of their other data sets, or I'll just reconstruct the daily data from the hourly OBS. Not sure I'm going to do that today, I haven't bought any gifts yet.

This is why I use a heavily weighted average of COOP stations within a certain distance (approximately 50 km). Of course, if the target station is "wacky" and not missing then the averaging doesn't totally fix it due to how heavily weighted the target is. I usually like to make it so the target has a weight of 1/4 to 1/2 the others combined... usually that gives it a weight of 20 compared to 1-10 for the others depending on their distance. The reconstruction caught some things that the Tulsa station data didn't due to all the "9999" sections that the NWS probably counted as "0" for snowfall when adding up their totals. Basically, the 1910's were a lot snowier than the NWS Tulsa Climatology Page makes them out to be.

[Rainshadow]

I think I've compiled this data for you, and everyone else...

At the bottom of this post there are two links to Excel workbooks (.xlsx type). The first Excel workbook, Discrete Snowfall Events, has all the >.1" snowfalls for each of RIC, IAD, DCA, BWI, PHL, NYC and BOS along with the AAM data as show here. Each event is listed by date along with the daily high/low, total liquid precip, snowfall, snow cover at the observation time, and finally the five AAM data points from the previously mentioned site.

The second workbook, Grouped Snowfalls, is the result of running the "discrete events" through a script that aggregates them into "storms". This is an attempt to show consecutive, multi-day accumulations as a single storm. I used a threshold of four days as the constraint for "longest storm duration". The first three columns reference the first date of measurable snow, the "peak date" (IOW, day with the highest accumulation in the set of days), and the last date of measurable snow in the series. The total precip is the sum of all the liquid measurements while the total snow is, obviously, the sum of the snow measurements. The AAM values should correspond to the value found on the "PEAK DATE".

Some caveats:

• I've only spot checked the values for each of the locations. There could be mistakes. Just let me know
  
• The beginning of time is 1/1/1900, so Excel likes to put observations from the 1800s at the bottom of an "oldest -> newest" sort.
  
• A lot of the pre 1960s snow cover measurements are pretty screwy. Either that or the script I used to originally pull those values is screwy. I haven't bothered to fix those errors and I probably won't. No, NYC didn't really have a 124" depth measurement.
  
• There are real gaps in some of the data. Those are shown as blanks and/or "M".
  

Discrete Snow Events

Grouped Snowfalls

Thank-you so much for sharing.

[Ender]

This is why I use a heavily weighted average of COOP stations within a certain distance (approximately 50 km). Of course, if the target station is "wacky" and not missing then the averaging doesn't totally fix it due to how heavily weighted the target is. I usually like to make it so the target has a weight of 1/4 to 1/2 the others combined... usually that gives it a weight of 20 compared to 1-10 for the others depending on their distance. The reconstruction caught some things that the Tulsa station data didn't due to all the "9999" sections that the NWS probably counted as "0" for snowfall when adding up their totals. Basically, the 1910's were a lot snowier than the NWS Tulsa Climatology Page makes them out to be.

I do this to QA the data. Basically I build a table of all the stations in a region and their line of sight distance, and elevation delta, from the station in question. Then, as I'm processing the data for a station, I look for "abnormalities" things like temperatures that are more than 1.5 SD removed from the mean, snowfall or cover measurements that "don't fit their climo", etc. Then, when I find such cases, I pull the hourly observations for that station and surrounding stations, as well as those surrounding station's observed Hi/Lo/precip for the day, and essentially check the fit of the observation in question. While automated this is a somewhat tedious and slow process. Not to mention the phenomenally large hashes that get built (I work through a state at a time, but that usually means loading the surrounding states as well). If I find that "observation doesn't fit" I retain the observation in its appropriate column, but I also create a new observation, in the separate columns, along with a flag explaining why the original observation was questionable.

Thank-you so much for sharing.

You're welcome. I later added region 3.4 temperatures if you'd like the sheet with that column as well just PM me.

[Rib]

HM,

Thanks. I still worry though. On the Euro its still nearly 144 hours away (again). I'm not going to guess what phase the gwo is going to be come December 26th-27th (I'd be wrong anyway), if it so happens to still be in the "southern hemisphere" with an aam < -1 (I left the chart at work) and Philly gets 10" or greater, it would be a first since 1958 proving there are exceptions to every rule. Before it made this left turn it was heading toward the phases it was in during 1/96 & 1/2000.

It seems the past couple of years, PHL is making foot+ storms look like moderate events I don't remember a stretch like this with several foot+ storms within such a short period of time.

[SnowGoose69]

It seems the past couple of years, PHL is making foot+ storms look like moderate events I don't remember a stretch like this with several foot+ storms within such a short period of time.

Its gonna suck starting in like 2030 when we go back into the +NAO state again, of course its hard to imagine its going to be as awful as the 1980-2000 +NAO state was given the previous +ERA before the 60s while less snowy was not that bad.