Jump to content
  • Member Statistics

    17,608
    Total Members
    7,904
    Most Online
    Chargers10
    Newest Member
    Chargers10
    Joined

January 2012 Storm Threats and Discussion


PhineasC

Recommended Posts

I am almost the last person on this board to discuss long range pattern changes and 10-day hopes, so I won't. But I am so desperate for anything to track that I looked at the 6z GFS just out. And there is not much to track. It has the 156 hour or so storm that drives up toward the eastern great lakes and makes everyone on the east coast very wet. Beyond that, it turns sharply colder as advertised on other models, and tries to give the hint of something in GOM, but crushes it well to the south and sends it easily out to sea.

I think.

Not sure if that look is a good look with regard to the hoped for pattern change or not, but cold and dry while precip is crushed to our south and tossed well out to sea would be a new way to not get snow this winter.

The GFS... crushing storms to the south in the long range? Oh and did you say the 6z GFS? If one is to attempt to interpret the models they should have an idea of the biases as well as the difficulties of resolving details in the long range, esp with changing patterns. I think maybe you had it right when you said you shouldnt discuss pattern changes ;)

Link to comment
Share on other sites

  • Replies 1.7k
  • Created
  • Last Reply
The GFS... crushing storms to the south in the long range? Oh and did you say the 6z GFS? If one is to attempt to interpret the models they should have an idea of the biases as well as the difficulties of resolving details in the long range, esp with changing patterns. I think maybe you had it right when you said you shouldnt discuss pattern changes ;)

I seriously want to know why we even have a 06z GFS run. It is horrible and usually the opposite of what all the other models show. Get rid of it or improve it.

Link to comment
Share on other sites

i came across this and was surprised to see the gfs in dead last out of the all models. this was from 144 hours out at 500mb for 0z 1.3.12

post-4-0-01022800-1325649556.jpg

here is all the euro and gfs model runs compared for each cycle, and yes it looks like the 6z and 18z are the worst runs

post-4-0-29636300-1325649786.jpg

Essentially the 00Z, 06Z and 18Z GFS scores are statistically a dead heat. When you have to take correlations scores to the third digit to find a difference the differences are statistically insignificant.

Link to comment
Share on other sites

94 was the absolute worst winter I can ever remember. Bitterly cold and nothing but Ice. It almost equalled the winter of 77 when the chesapeake bay froze so bad they had to bring in coast guard ice breakers to free shipping

Lanes. 77 had very little snow to my recollection.

I agree. I was working as a busser at a restaurant in Fells Point and the restaurant shut down when pipes froze and flooded (and iced in some spots) the place. As for the Winter of '77 I turned one in June of '77 so I obviously don't remember a thing about it.

Link to comment
Share on other sites

I seriously want to know why we even have a 06z GFS run. It is horrible and usually the opposite of what all the other models show. Get rid of it or improve it.

From a met? really? :axe:

Also, even if we didn't make those forecasts public, we would still need to run the 06z GFS to drive all of the down stream models (NAM, HWRF, GFDL, Ruc/Rapid Refresh, Hycom, etc.).

Seriously, please don't post about models if you don't know what you're talking about.

** Edit to add: Even if a 120h forecast from 6z verifies slightly worse than the 120h forecast from 0z, recall that the two points for verification are not the same. The 114h forecast from 6z is almost always better than the 120h forecast from 0z, for example. If no data were assimilated at 6z, the 120h forecast from 0z would be identical to the 114h forecast from 6z.

Link to comment
Share on other sites

From a met? really? :axe: Also, even if we didn't make those forecasts public, we would still need to run the 06z GFS to drive all of the down stream models (NAM, HWRF, GFDL, Ruc/Rapid Refresh, Hycom, etc.). Seriously, please don't post about models if you don't know what you're talking about.

Yes from a met. And instead of being a jackass with your answer, maybe you could answer more appropriately.

Mets still learn as well. Not all mets understand all the details that go into models. You can actually be a pretty damn good forecaster without knowing that, which I am and was, so don't tell me not to post. And all of what you told me isn't exactly taught in school.

All I was saying is a lot of people don't even bother looking at the 06z and 18z GFS because they're usually not as accurate. And I didn't understand why, but thanks for your pathetic answer.

Please don't post if you're gonna answer questions in the ways you do.

And in response to your edited version to your post. Yes I know that about the models and their hours. THAT, a met should know.

Link to comment
Share on other sites

I seriously want to know why we even have a 06z GFS run. It is horrible and usually the opposite of what all the other models show. Get rid of it or improve it.

Yes from a met. And instead of being a jackass with your answer, maybe you could answer more appropriately.

Mets still learn as well. Not all mets understand all the details that go into models. You can actually be a pretty damn good forecaster without knowing that, which I am and was, so don't tell me not to post. And all of what you told me isn't exactly taught in school.

All I was saying is a lot of people don't even bother looking at the 06z and 18z GFS because they're usually not as accurate. And I didn't understand why, but thanks for your pathetic answer.

Please don't post if you're gonna answer questions in the ways you do.

And in response to your edited version to your post. Yes I know that about the models and their hours. THAT, a met should know.

I believe I answered the bolded above, even if you completely ignore the perceived lack of skill for the referenced cycle. I'm not about to get into a pissing contest over something so stupid. Given that you called it "horrible", "the opposite of what other models show", and to "get rid of it", I don't think my response was that out of line (sorry for my negative tone, however). I shouldn't have said "not to post about models", so I apologize (though, again, your original post was pretty ridiculous in my opinion).

This issue of 6z/18z versus the other cycles, perceived skill, etc., has been discussed many times in many threads. If you have an actual question or want to learn more, then ask. However, if my replies are so offensive, I'm happy to keep my thoughts to myself.

Link to comment
Share on other sites

Yes from a met. And instead of being a jackass with your answer, maybe you could answer more appropriately.

Mets still learn as well. Not all mets understand all the details that go into models. You can actually be a pretty damn good forecaster without knowing that, which I am and was, so don't tell me not to post. And all of what you told me isn't exactly taught in school.

All I was saying is a lot of people don't even bother looking at the 06z and 18z GFS because they're usually not as accurate. And I didn't understand why, but thanks for your pathetic answer.

Please don't post if you're gonna answer questions in the ways you do.

And in response to your edited version to your post. Yes I know that about the models and their hours. THAT, a met should know.

Perhaps you, as a met, should not throw out such lines as "Get rid of it or improve it" when not knowing what goes into the model. This is especially true when you have a developer on the forum who is able to and does provide the correct information in addition to discussing some of the finer points and future plans.

His tone is fine... he has addressed the depreciation of accuracy of the off-hour runs several times in this sub-forum and has dispelled the myth that the off-hour runs are garbage, with the stats to back it up. It's insulting to him and the other developers to have to repeatedly get these criticisms and false assumptions.

-----

EDIT: Also, what dtk said above (he posted while I was writing)

Link to comment
Share on other sites

Now experiencing my 60th winter, the great bulk of which were in MD/DC/DE, I can say without hesitation the winter of '76-'77 stands unparalleled.

Some winters are snowy, some not; some winters cold, some not; but I have to wonder just how great are the statistical odds against three consecutive months in D.C. running 10F below normal.

January 1977 was remarkably cold in DC, with an average temperature of 25.4 degrees, ranking it as the sixth coldest month ever -- and the coldest since January 1940. However, December 1976 (average temperature of 35.5) and February 1977 (average temperature of 38.8) weren't all that bad. The most remarkable thing about January 1977 was that, for 20 consecutive days, the temperature never got into the 40s, with the high temperature range being 18-39 and the low temperature range being 2-27 during that period (mean of 23.6).

P.S. More relevant to this topic, AccuWeather currently shows an East Coast snowstorm on Martin Luther King Day, January 16.

Link to comment
Share on other sites

Yes from a met. And instead of being a jackass with your answer, maybe you could answer more appropriately.

Mets still learn as well. Not all mets understand all the details that go into models. You can actually be a pretty damn good forecaster without knowing that, which I am and was, so don't tell me not to post. And all of what you told me isn't exactly taught in school.

All I was saying is a lot of people don't even bother looking at the 06z and 18z GFS because they're usually not as accurate. And I didn't understand why, but thanks for your pathetic answer.

Please don't post if you're gonna answer questions in the ways you do.

And in response to your edited version to your post. Yes I know that about the models and their hours. THAT, a met should know.

Any people who aren't using off hour model runs are making a mistake. I also think DTK answer was appropriate given your outlandish original statement. Maybe he shouldn't have made the comment about coming from a met but I'm sure that was out of frustration as the myth that the 06Z and 18Z model runs are way worse than the 00Z or 12s runs. that myth has been discussed with verification stats numerous times on this board. If you go back to midlo's post in this thread and look at the difference in anomaly correlation scores. You'll see that you have to to to the third decimal point to order them. That is not a statistically significant difference. Essentially, the verification suggests little difference in the runs.

Link to comment
Share on other sites

Perhaps you, as a met, should not throw out such lines as "Get rid of it or improve it" when not knowing what goes into the model. This is especially true when you have a developer on the forum who is able to and does provide the correct information in addition to discussing some of the finer points and future plans.

His tone is fine... he has addressed the depreciation of accuracy of the off-hour runs several times in this sub-forum and has dispelled the myth that the off-hour runs are garbage, with the stats to back it up. It's insulting to him and the other developers to have to repeatedly get these criticisms and false assumptions.

-----

EDIT: Also, what dtk said above (he posted while I was writing)

The funny thing is I started digging some more into the past 60 days worth of scores (various metrics), and the 06z cycle does seem to be systematically slightly worse than the other cycles (particularly over the NH/Conus)...though the differences aren't statistically significant (at least not at a 95% confidence level) for the same lead times. However, the differences aren't even close to a 6h advantage (meaning, as I said before, a 114h forecast from 6z will almost always be better than a 120h forecast from 0z.....and those are the two forecasts valid at the same time....so it's not degrading anything, per say).

As an aside, I don't think that I (nor my colleagues) mind criticism if/when it's warranted.

p.s. This winter sucks. I really need it to snow > 1" on the 16th so that I can win our office "first snowfall pool".

Link to comment
Share on other sites

Yes from a met. And instead of being a jackass with your answer, maybe you could answer more appropriately.

Mets still learn as well. Not all mets understand all the details that go into models. You can actually be a pretty damn good forecaster without knowing that, which I am and was, so don't tell me not to post. And all of what you told me isn't exactly taught in school.

All I was saying is a lot of people don't even bother looking at the 06z and 18z GFS because they're usually not as accurate. And I didn't understand why, but thanks for your pathetic answer.

Please don't post if you're gonna answer questions in the ways you do.

And in response to your edited version to your post. Yes I know that about the models and their hours. THAT, a met should know.

I remember lots of people in met school didn't take the 6z or 18z runs of the ETA/AVN seriously, but I think that thought process is somewhat outdated. There have been multiple trends that have begun on intermediate runs of those models. Cannot remember exact storms at this moment.

Link to comment
Share on other sites

Wes it will be...not foolish is he with forecasting emotion...facts he deals in not wishes.

he is my Yoda

Oh please, I can beat Wes on the golf course, and I'll take him down here too :whistle: . Oh wait, he was already sort of on my side....

Link to comment
Share on other sites

Wes it will be...not foolish is he with forecasting emotion...facts he deals in not wishes.

he is my Yoda

The only thing I could think of when I saw Wes's post was

FINISH HIM

... snowdude is a good guy though, so I don't wanna knock him too much

Link to comment
Share on other sites

Oh please, I can beat Wes on the golf course, and I'll take him down here too :whistle: . Oh wait, he was already sort of on my side....

I think I did tie you or beat you once using Buzz's rules which meant I really lost to you that time too. I recently shot a 38 on the front nine at Twin Shields and only used two mulligans though I now play from the yellow tees. So just wait, next time I'll crush you like a bug.

Link to comment
Share on other sites

As an aside, I don't think that I (nor my colleagues) mind criticism if/when it's warranted.

p.s. This winter sucks. I really need it to snow > 1" on the 16th so that I can win our office "first snowfall pool".

Bolded for emphasis :)

I lost my snow pool :( Was hoping for late Nov/early Dec to pay off in an early blocking pattern that never developed.

Link to comment
Share on other sites

I think I did tie you or beat you once using Buzz's rules which meant I really lost to you that time too. I recently shot a 38 on the front nine at Twin Shields and only used two mulligans though I now play from the yellow tees. So just wait, next time I'll crush you like a bug.

Well played sir, well played. I think I need to abide by Buzz's "rules" from now on, it will make golf a lot more fun.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.

×
×
  • Create New...