January 2013 mid-long range disco thread

[Deck Pic]

Cheap shot.  If I remember he went very mild for 3 months, 2 out of 3 were, the season was slightly above.  That's not a bust and the words "great outlook" were very likely never used.

 

I will say this about you and he though. Over last 10 years you two have the overall best track records of anyone and anything.

If any entity was dependent upon both of your seasonal outlooks and your assessment with winter persistence patterns then they would be far better off than all the NOAA, CPC, NWS seasonal outlooks and pattern recognition.  I have for years championed the analog method employed by Keith and you and even though I highly believe in both of your skills I think it does damn the "official" products we are currently stuck with when you consider that two non-degreed mets, essentially very skilled hobbyists, can outforecast all the outfits that cost millions and billion per year to operate. 

 

You're right...it was a bit of low blow...It wasn't a bad outlook nor a great one.....I think seasonal forecasting is in its infancy still and not a lot of money or resources are put into it because of the low confidence..I do agree with you about the models to SOME extent....These products are made available to the public without any sort of objective guidelines...It is common sense that they will get worse the further out you go....I am perfectly fine with the fact that they don't "work" when you get far out..I think what they can do is amazing....I don't expect any algorithm to be skilled enough to have any sort of precision outside of a few days with something so chaotic and complex with so many moving parts as weather patterns..I think there are plenty of people who have the skill and experience to understand how to use the models, when not to use them, what purpose to use them for and their biases and limitations...and as a result use probabilistic forecasting which to me is the best type of forecasting....I am slightly bothered, however, that the government releases a product to the general public without any sort of manual or guidelines that essentially doesn't work outside a certain range and cannot be taken at face value...

 

As far as me and KA...I think we do better than a coin flip when you look at the body of work...generally B- average versus a C which you would get from a coin flip....I try to add color to the CWG outlooks because I think just throwing out numbers is a tall task in the mid atlantic where we are almost never in the core of cold or warm anomalies and often caught in between....I still think there is a folksy/entertainment value aspect to these which is fine as well and we still need to broad brush....What I think is equally important is I have a 9 year body of work and KA a 12 year available body of work with actual numbers that can be assessed by anyone...There are a lot of naysayers but I honestly do not know of ANYONE else who when asked to link to their say last 5 outlooks will or is able to do so....

[WEATHER53]

You're right...it was a bit of low blow...It wasn't a bad outlook nor a great one.....I think seasonal forecasting is in its infancy still and not a lot of money or resources are put into it because of the low confidence..I do agree with you about the models to SOME extent....These products are made available to the public without any sort of objective guidelines...It is common sense that they will get worse the further out you go....I am perfectly fine with the fact that they don't "work" when you get far out..I think what they can do is amazing....I don't expect any algorithm to be skilled enough to have any sort of precision outside of a few days with something so chaotic and complex with so many moving parts as weather patterns..I think there are plenty of people who have the skill and experience to understand how to use the models, when not to use them, what purpose to use them for and their biases and limitations...and as a result use probabilistic forecasting which to me is the best type of forecasting....I am slightly bothered, however, that the government releases a product to the general public without any sort of manual or guidelines that essentially doesn't work outside a certain range and cannot be taken at face value...

The only "what should the models do" suggestion that I can offer is that they are probalby too finely meshed. Kinda like looking at the Statue of Liberty with a microscope-you will get some intricate detail but miss the big picture.

[BristowWx]

GFS ens members would say don't give up on next week just yet.

Hr 240 looks not too bad...no reason to jump ship...our expectations are low anyway so heck with it...

[Deck Pic]

The only "what should the models do" suggestion that I can offer is that they are probalby too finely meshed. Kinda like looking at the Statue of Liberty with a microscope-you will get some intricate detail but miss the big picture.

 

I kind of agree...I'm not entirely sure of the value of a day 14 product that is released in the exact same format as a day 1

[Bob Chill]

..I am slightly bothered, however, that the government releases a product to the general public without any sort of manual or guidelines that essentially doesn't work outside a certain range and cannot be taken at face value...

 

Even if they did nobody would read it. Model panels are like watching a video to most. They just look at it and think it's the most likely outcome. I have a feeling most folks only look @ 850/precip and 2m panels. Just like looking at a radar loop.

 

I learned years ago that if you don't understand 500 maps then don't even bother looking at a model and thinking you know what's going on. 

 

Another thing about the natural swings in the lr is few if any take the time to put the last 2-3 runs side by side to see where the major difference is @ 500. This simple skill really helps a hobbyist get a handle on how to track and follow the changes and making at the very least semi-educated decisions on which way things are "leaning" in the lr. 

 

I couldn't agree more with your praise about how well the models do nowadays. All too many folks focus on a microscopic piece of real estate on a global model in the lr and yell "models suck" and things like that. My goodness that is unrealistic. 

 

The best course outside of d5-6 is to simply key on the longwave features. Understanding the 500mb polar panels on the gfs are an outstanding way of getting a handle on what to expect @ mid-latitudes. We rely on the setup to our north more than to our west when making sense of longer leads. 

 

As for the latest trends in the op runs, there is no reason to doubt the retrograding trough/ridge setup in the pac. We've already been here. On the other hand, having the pv on our side can change things in a hurry too. Even though the op runs are disheartening, I'm not sweating it yet because a lot has to transpire first before we freak out on the pac setup. Just look at the ens forecasts for the major indicies. Quite a bit of spread. Nothing is set in stone. I was never really confident in a major cold outbreak anyways because we've seen it time and time again in the lr just to have it not materialize. Hardly a shocking turn of events if it doesn't happen this time.

[WinterWxLuvr]

Even if they did nobody would read it. Model panels are like watching a video to most. They just look at it and think it's the most likely outcome. I have a feeling most folks only look @ 850/precip and 2m panels. Just like looking at a radar loop.

 

I learned years ago that if you don't understand 500 maps then don't even bother looking at a model and thinking you know what's going on. 

 

Another thing about the natural swings in the lr is few if any take the time to put the last 2-3 runs side by side to see where the major difference is @ 500. This simple skill really helps a hobbyist get a handle on how to track and follow the changes and making at the very least semi-educated decisions on which way things are "leaning" in the lr. 

 

I couldn't agree more with your praise about how well the models do nowadays. All too many folks focus on a microscopic piece of real estate on a global model in the lr and yell "models suck" and things like that. My goodness that is unrealistic. 

 

The best course outside of d5-6 is to simply key on the longwave features. Understanding the 500mb polar panels on the gfs are an outstanding way of getting a handle on what to expect @ mid-latitudes. We rely on the setup to our north more than to our west when making sense of longer leads. 

 

As for the latest trends in the op runs, there is no reason to doubt the retrograding trough/ridge setup in the pac. We've already been here. On the other hand, having the pv on our side can change things in a hurry too. Even though the op runs are disheartening, I'm not sweating it yet because a lot has to transpire first before we freak out on the pac setup. Just look at the ens forecasts for the major indicies. Quite a bit of spread. Nothing is set in stone. I was never really confident in a major cold outbreak anyways because we've seen it time and time again in the lr just to have it not materialize. Hardly a shocking turn of events if it doesn't happen this time.

 

All modeled or dependent on models.  Using ensembles to predict 11-15 days, same thing.  The underlying assumption when you look at patterns, 500 maps, and on and on and on, is that you are looking at modeled patterns, 500 maps, etc.  None of them have shown themselves to be reliable.  I don't know who the poster was earlier who said it, but I think 7-10 forecasting based on climo and analogs would trump modeled forecasts,  wrt the upper air pattern.  

 

The positive to take from it is that if the 11-15 day outlook isn't what you want, you have a good chance that it won't be that way anyway.

 

I'm not sure I buy into the whole "the atmosphere has a memory" idea either.  This could change any minute.  1989-1990 is a perfect example of a total flip.  I'm not saying that that year has anything in common with this one, just that it can change completely, quickly.

 

Whatever the top 10 analogs, combined with SST data, suggest for the forecast 7-10 days from now, is probably, IMO, the most likely outcome.  It would be interesting if someone could post that 500 map (day 7, and day 10) using that technique, and then compare it to what happens at those times.

[WxUSAF]

Good look on the Euro day 9-10 (+PNA, PV over SE Canada, -NAO, southern stream)...but it's still day 9-10. 

[BristowWx]

Good look on the Euro day 9-10 (+PNA, PV over SE Canada, -NAO, southern stream)...but it's still day 9-10. 

Looks like 480 thickness north of Maine

[usedtobe]

The 12Z GEFS ensmeble are much slower with the retrogression which would widen the window of how long we might have a chance to get lucky before any warm up. As I tried to note in my earlier assessment is that this is a case where even the bigger features are not being forecast that consistently, especially the vortex over Hudson bay which on some runs ends up too far north of us and on others ends up far enough southeast in Canada to give us a chance for something. That feature is really important and even with a less than stellar Pacific could keep enough confluence around to get some winter weather if the timign were right on a shortwave. Also note how different the Euro and GFS are in handling the southern stream shortwave, the gfs seems to bury it way out west while the operational Euro shears it eastward. The euro hadling of the hudson basy vortex seems pretty consistent with the 12Z GEFS 240 hrs handling and is a much better look for us than the 12Z operational GFS which looks to be an outlier. I'm still not sure what to make of the pattern and think in this case Winterwxluvr's comments about the long range are pretty good. Last week the 14 day forecast was a no brainer that we'd end up torching this weak. This week the signals are much more mixed.

[Bob Chill]

The models have been reliable overall with longwave stuff. I don't know how anybody can say otherwise. I think it's way too easy to get hung up on a "nice couple of runs" and forget all the crappy ones before and then complain about the ones after. Models killed it in december. Especially the first 2-3 weeks.  They've also done an excellent job so far this month. Pattern changes simply cannot be accurately modeled. Timing the breakdown of a persistent pattern and showing where it's going 15 days in advance is probably mathematically impossible. 

 

As soon as we see things in the lr bouncing all over for 2-3+ days then we know we are in for some sort of sizeable pattern change. Whenever that starts to happen, posters like Wes and Zwyts and Co do a nice job deciphering and coming up with the most likely outcome. It good to take a step back from run to run verbatim roller coaster and look at the big picture. 

 

I personally hope I'm wrong with my opinion but I'm leaning towards a cool but not cold period with a messy storm track and then I would place a very small bet on the -pna re-establishing itself. I want to be wrong  and my opinion doesn't mean much anyways but it's how I'm feeling based on what I'm seeing. 

 

There is a bit of a recurring theme this winter in the pac and it's not a coincidence so there are factors that I wouldn't call "memory" but rather "normal" or "expected" given the fact that we were all banking on a Nino that vanished. I can't even count the # of posts in Mitch's cfs thread about how if the nino doesn't happen then we're fooked. Zwyts in particular was very precise with his posts irt to enso state.

 

For all intents and purposes we are basically in a weak nina enso. Or Nina hangover or whatever you want to call it.  The pattern sure resembles one. The persistent -pna is part of the hand we are dealt. There are some good years in the past that have similar features irt the state of the pna and enso. January of 57, 72, 90, and 02 all have very similar characteristics to what we are experiencing. 

 

These 500 anoms in Jan for those years look awful familiar. The ridge placement in the pac is undeniably consistent. I don't see how this month ends up vastly different at this time. 

 

 

 

 

 

 

 

 

 

 

[Swiscaster]

It's funny how we are talking about the difficulties in long-range weather forecasting when it is currently 43 degrees and raining at Dulles. 16 hours ago, the NWS forecast was "58 and Mostly Sunny" for this afternoon.

[usedtobe]

It's funny how we are talking about the difficulties in long-range weather forecasting when it is currently 43 degrees and raining at Dulles. 16 hours ago, the NWS forecast was "58 and Mostly Sunny" for this afternoon.

and I'm 58 so the for me the forecast wasn't that bad in mby though I am mostly cloudy. It's a strong warm advection pattern which usually leads to at least some clouds.

[Deck Pic]

The 12Z GEFS ensmeble are much slower with the retrogression which would widen the window of how long we might have a chance to get lucky before any warm up. As I tried to note in my earlier assessment is that this is a case where even the bigger features are not being forecast that consistently, especially the vortex over Hudson bay which on some runs ends up too far north of us and on others ends up far enough southeast in Canada to give us a chance for something. That feature is really important and even with a less than stellar Pacific could keep enough confluence around to get some winter weather if the timign were right on a shortwave. Also note how different the Euro and GFS are in handling the southern stream shortwave, the gfs seems to bury it way out west while the operational Euro shears it eastward. The euro hadling of the hudson basy vortex seems pretty consistent with the 12Z GEFS 240 hrs handling and is a much better look for us than the 12Z operational GFS which looks to be an outlier. I'm still not sure what to make of the pattern and think in this case Winterwxluvr's comments about the long range are pretty good. Last week the 14 day forecast was a no brainer that we'd end up torching this weak. This week the signals are much more mixed.

 

yeah...Euro looks nice at day 10

[snowfan]

You're at 58, I'm at 48. 10 degree difference over maybe 28-30 miles as the crow flies.

and I'm 58 so the for me the forecast wasn't that bad in mby though I am mostly cloudy. It's a strong warm advection pattern which usually leads to at least some clouds.

 

 

[mitchnick]

Accuwx maps haven't updated but day 8'ish looks sort of interesting on the 12Z Euro

anything falling from that -2 or -3C 850 temps for those with storm vista?

[WinterWxLuvr]

The models have been reliable overall with longwave stuff. I don't know how anybody can say otherwise. I think it's way too easy to get hung up on a "nice couple of runs" and forget all the crappy ones before and then complain about the ones after. Models killed it in december. Especially the first 2-3 weeks.  They've also done an excellent job so far this month. Pattern changes simply cannot be accurately modeled. Timing the breakdown of a persistent pattern and showing where it's going 15 days in advance is probably mathematically impossible. 

 

As soon as we see things in the lr bouncing all over for 2-3+ days then we know we are in for some sort of sizeable pattern change. Whenever that starts to happen, posters like Wes and Zwyts and Co do a nice job deciphering and coming up with the most likely outcome. It good to take a step back from run to run verbatim roller coaster and look at the big picture. 

 

I personally hope I'm wrong with my opinion but I'm leaning towards a cool but not cold period with a messy storm track and then I would place a very small bet on the -pna re-establishing itself. I want to be wrong  and my opinion doesn't mean much anyways but it's how I'm feeling based on what I'm seeing. 

 

There is a bit of a recurring theme this winter in the pac and it's not a coincidence so there are factors that I wouldn't call "memory" but rather "normal" or "expected" given the fact that we were all banking on a Nino that vanished. I can't even count the # of posts in Mitch's cfs thread about how if the nino doesn't happen then we're fooked. Zwyts in particular was very precise with his posts irt to enso state.

 

For all intents and purposes we are basically in a weak nina enso. Or Nina hangover or whatever you want to call it.  The pattern sure resembles one. The persistent -pna is part of the hand we are dealt. There are some good years in the past that have similar features irt the state of the pna and enso. January of 57, 72, 90, and 02 all have very similar characteristics to what we are experiencing. 

 

These 500 anoms in Jan for those years look awful familiar. The ridge placement in the pac is undeniably consistent. I don't see how this month ends up vastly different at this time. 

 

That's a very valid way to look at it, IMO. 

[Ian]

Accuwx maps haven't updated but day 8'ish looks sort of interesting on the 12Z Euro

anything falling from that -2 or -3C 850 temps for those with storm vista?

 

nope

[Midlo Snow Maker]

yeah...Euro looks nice at day 10

ens. look great, plenty cold and has precip......... atleast until 0z

[WinterWxLuvr]

ens. look great, plenty cold and has precip......... atleast until 0z

Is this in reference to the 144-168 time frame or more along the day 9-10 range?

[Ji]

Is this in reference to the 144-168 time frame or more along the day 9-10 range?

 

really? our snow threats are always in the 9-10 range. the day 10 storm yesterday on the 12z euro is still a 10 day storm except now surpressed

[mitchnick]

really? our snow threats are always in the 9-10 range. the day 10 storm yesterday on the 12z euro is still a 10 day storm except now surpressed

 

at 10 days, would you prefer a) hit, b ) miss to the north, c) miss to the west, or d) miss to the south

 

I'll pick d) at 10 days

[Bob Chill]

Mitch, you forgot E ) none of the above because their is no storm. 

[WinterWxLuvr]

Mitch, you forgot E ) none of the above because their is no storm.

Just a matter of time. If that ec ens pattern at 10 days couldn't produce something of interest, I give up. But....it is 10 days. So .....

[mitchnick]

Mitch, you forgot E ) none of the above because their is no storm. 

 

I gave him a choice and wanted him to pick the best answer, not the right one  

 

btw, the ensembles do have yesterday's 12z storm going off the SC coast at 228 hrs and it's plenty cold (too cold thanks to the 50/50)

[Bob Chill]

On a more serious note....

 

I've been watching the mjo forecast for over a week once it decided to show a pretty strong wave propagating. Don't read into the cheese puff curl at the end of guidance too much. It's been there for days but keeps moving forward through the octants as time goes by. When the mjo was first signaled to be active it was actually near the phase 2/3 border at onset but it emerged out of the COD at phase 3/4 border. 

 

The GEFS didn't really show it progressing like it is either. It was more of a slow and erratic motion. It's moving with authority now. The disco in the sne thread is really good if anyone is interested in the more technical side of what's going on. 

 

At the current pace the index could be firmly in phase 7 within a week. My hunch is that the fade and curl at the end of guidance will continue to be pushed forward. I've checked the index daily for a while and every day shows a further push towards "friendly" territory with no signs of changing course over the first 7 days of each run of the GEFS.

 

 

 

Here's the JFM temp composites for phases 6-7-8:

 

 

 

It's just a piece of the puzzle to me and I'm just a beginner irt understanding what it all means for us. There is much uncertainty in guidance once we go beyond d7-8 so anything can happen. At the very least, it's nice to see the mjo become active and send a strong signal. Phase 6 temp composite really lines up well with what is progged at the same time for the eastern half of the country. 

 

We are getting teased with some real below normal temps as it is in the lr but it's prob a coin flip at best right now. If the mjo keeps up the strength it currently has AND cruises through the 8-1-2 promised land than I would hedge towards us getting some decent cold here for an acceptable amount of time (whatever that means). 

[WEATHER53]

That's a very valid way to look at it, IMO. 

 

And I agree also. And if we can see that, and the results that have followed and what is looming, then why can't the million dollar babies see it.  I bet there is a set of data, even if small, that would include a projected nino following a nina that failed to become a nino- and then what followed?

[Deck Pic]

And I agree also. And if we can see that, and the results that have followed and what is looming, then why can't the million dollar babies see it.  I bet there is a set of data, even if small, that would include a projected nino following a nina that failed to become a nino- and then what followed?

 

there are products out there that use analogs..

 

it all depends on what you are trying to accomplish....I think you need to combine a lot of factors..an analog from say 1952 is not going to tell you what happens next week...But I think it is good to recognize that patterns can have certain default states...we are in a -PDO regime with a Nina hangover and this pattern has driven the weather..and gives us bad storm tracks....so far the PAC pattern  has dominated, but that may change as we get into better climo...-AO/-NAO blocking is probably more of a player for us during better climo and can counteract the pacific better...

 

But yes...I think an understanding of past analogs and climatology and persistence can help give us pause whenever we see the models try to do something that goes against our default pattern....

 

I think those that combine all of these elements give the best assessments and forecasts...Wes has been doing it all winter...when the models are outside their useful range as far as any specifics, assess whether the longwave pattern supports the ideas the models are trying to convey...this is tricky stuff.....no method is going to allow you to get the details right...all we could say in Dec is that the colder pattern still supported a storm track to our west...but that doesnt mean it couldnt go to our south...or what cold air may be in place...or where the rain snow line will be...analogs aren't going to tell us if the low will be far enough east to give Leesburg snow vs rain...

[Deck Pic]

All modeled or dependent on models.  Using ensembles to predict 11-15 days, same thing.  The underlying assumption when you look at patterns, 500 maps, and on and on and on, is that you are looking at modeled patterns, 500 maps, etc.  None of them have shown themselves to be reliable.  I don't know who the poster was earlier who said it, but I think 7-10 forecasting based on climo and analogs would trump modeled forecasts,  wrt the upper air pattern.  

 

The positive to take from it is that if the 11-15 day outlook isn't what you want, you have a good chance that it won't be that way anyway.

 

I'm not sure I buy into the whole "the atmosphere has a memory" idea either.  This could change any minute.  1989-1990 is a perfect example of a total flip.  I'm not saying that that year has anything in common with this one, just that it can change completely, quickly.

 

Whatever the top 10 analogs, combined with SST data, suggest for the forecast 7-10 days from now, is probably, IMO, the most likely outcome.  It would be interesting if someone could post that 500 map (day 7, and day 10) using that technique, and then compare it to what happens at those times.

 

 rolling  over analogs is a pretty bad method for short term patterns...We do it because it is interesting but it doesnt work very well...analogs are good for broadbrushing and giving us mean patterns over a decent sample size of days/weeks...usually a month or more.....The best thing we can do outside of 10 days is say we dont know...if we are going to tackle it since it is fun, Wes's probabilistic way is the best and if you have experience with models you can cull information....just like experience with anything else...we really dont know what kind of pattern we will be in 2 weeks from now nor do we need to know..But I think we can view guidance that wants to give us a stable -EPO/+PNA look with great skepticism....

[Amped]

On a more serious note....

 

I've been watching the mjo forecast for over a week once it decided to show a pretty strong wave propagating. Don't read into the cheese puff curl at the end of guidance too much. It's been there for days but keeps moving forward through the octants as time goes by. When the mjo was first signaled to be active it was actually near the phase 2/3 border at onset but it emerged out of the COD at phase 3/4 border. 

 

The GEFS didn't really show it progressing like it is either. It was more of a slow and erratic motion. It's moving with authority now. The disco in the sne thread is really good if anyone is interested in the more technical side of what's going on. 

 

At the current pace the index could be firmly in phase 7 within a week. My hunch is that the fade and curl at the end of guidance will continue to be pushed forward. I've checked the index daily for a while and every day shows a further push towards "friendly" territory with no signs of changing course over the first 7 days of each run of the GEFS.

 

mjo1.8.JPG

 

 

Here's the JFM temp composites for phases 6-7-8:

 

mjo678.JPG

 

 

It's just a piece of the puzzle to me and I'm just a beginner irt understanding what it all means for us. There is much uncertainty in guidance once we go beyond d7-8 so anything can happen. At the very least, it's nice to see the mjo become active and send a strong signal. Phase 6 temp composite really lines up well with what is progged at the same time for the eastern half of the country. 

 

We are getting teased with some real below normal temps as it is in the lr but it's prob a coin flip at best right now. If the mjo keeps up the strength it currently has AND cruises through the 8-1-2 promised land than I would hedge towards us getting some decent cold here for an acceptable amount of time (whatever that means). 

 

 

A strong MJO dying right on the phase 6-7 border is a good sign of a bad winter.

[Bob Chill]

A strong MJO dying right on the phase 6-7 border is a good sign of a bad winter.

 

lol- yes it is. But it's been dying like this @ 4-5 then 5-6 and most lately 6-7. The first 6-7 days of each run of guidance has been a bee line and the end has fizzled. But luckily it keep happening further into the octants every day. 

 

But your probably right. It will keep pushing forward until the beginning of phase 8 and then verify with a trip to the COD. It's just that kinda pattern...lol