Jump to content
  • Member Statistics

    17,797
    Total Members
    7,904
    Most Online
    manaja
    Newest Member
    manaja
    Joined

February Medium/Long Range Thread


stormtracker
 Share

Recommended Posts

3 minutes ago, Terpeast said:

That's certainly FOLKS worthy. Feb 2015 is looking like a decent analog at least for THIS run.

I was thinking the same but iirc that storm did a classic h5 close off in the deep south. It was a 1-2 punch because the closed ULL kinda dawdled. Dry slot lull was aggravating. CMC had the setup for a close off. Gfs and euro for most part setups would require a tpv phase for a bomb. That comes with excessive risk of a bust but would be fun AF for someone 

  • Like 2
Link to comment
Share on other sites

7 minutes ago, Terpeast said:

That's certainly FOLKS worthy. Feb 2015 is looking like a decent analog at least for THIS run.

Do you mean Feb 2014?  We had a lot of minor snows Feb 2015 but no amplified waves most were over running in the epo pattern. The biggest snow we had that month was that cutter that somehow produced 4-8” because it was so cold the day before. 
 

I see some similarities to the Feb 2014 storm. 

  • Like 1
Link to comment
Share on other sites

I've mostly lurked here over the past several years,  just posting my observations during snow events.  I have a couple of general questions about models, don't know if this is the right thread but don't want to post on a subforum that's dead with no chance of a response.

(1) When models spit out the contoured snow maps, do the figures within those maps represent a mean projected accumulation (i.e. if its 6 inches, that means a deviation of +/-  2" depending on confidence level), or do those totals represent the maximum projected accumulation (i.e. not more than 6", within confidence level of 95%). I know these maps are always a snapshot in time, and some maps are specific about what they're showing, such as the 10% chance/confidence maps.  But if not, are the figures just the mean?

(2) I often hearing that ______ model has a north bias, a south bias, generally projects higher qpf totals, etc.  If this is in fact true, meaning the bias is statistically significant, why aren't those models corrected to account for the bias? After all, aren't all of these models based on data or other markers from past storms, compared with incoming data from the current storm? And the purpose of any statistical model is to become progressively more accurate over time? Or are the people claiming real bias just weenies?

(3) Everyone seems to have slightly different opinions, but as far as model supremacy is concerned, how do each rank at various stages of a storm (168 hours, 98 hours, 48 hours, 24 hours, 12 hours).  Based on what I've observed, there seem to be three models that lead the pack, tell me if I'm right:

  • EURO-- Seems to be the consensus best model at all stages
  • GFS-- Seems to be about the same as the NAM, but
  • NAM-- Seems to given more attention at about the 24 hour stage
  • Then....  ICON, GEFS, UKIE, HRRR, GDEPS, GEM, CMA, etc.--  all works in progress.

Any input would be appreciated..  Thanks!!

  • Thanks 1
Link to comment
Share on other sites

3 minutes ago, grhqofb5 said:

I've mostly lurked here over the past several years,  just posting my observations during snow events.  I have a couple of general questions about models, don't know if this is the right thread but don't want to post on a subforum that's dead with no chance of a response.

(1) When models spit out the contoured snow maps, do the figures within those maps represent a mean projected accumulation (i.e. if its 6 inches, that means a deviation of +/-  2" depending on confidence level), or do those totals represent the maximum projected accumulation (i.e. not more than 6", within confidence level of 95%). I know these maps are always a snapshot in time, and some maps are specific about what they're showing, such as the 10% chance/confidence maps.  But if not, are the figures just the mean?

(2) I often hearing that ______ model has a north bias, a south bias, generally projects higher qpf totals, etc.  If this is in fact true, meaning the bias is statistically significant, why aren't those models corrected to account for the bias? After all, aren't all of these models based on data or other markers from past storms, compared with incoming data from the current storm? And the purpose of any statistical model is to become progressively more accurate over time? Or are the people claiming real bias just weenies?

(3) Everyone seems to have slightly different opinions, but as far as model supremacy is concerned, how do each rank at various stages of a storm (168 hours, 98 hours, 48 hours, 24 hours, 12 hours).  Based on what I've observed, there seem to be three models that lead the pack, tell me if I'm right:

  • EURO-- Seems to be the consensus best model at all stages
  • GFS-- Seems to be about the same as the NAM, but
  • NAM-- Seems to given more attention at about the 24 hour stage
  • Then....  ICON, GEFS, UKIE, HRRR, GDEPS, GEM, CMA, etc.--  all works in progress.

Any input would be appreciated..  Thanks!!

  1. if it's an ensemble map (GEPS, GEFS, EPS), it's usually a mean or median. If it's coming from the parent (the CMC, the GFS, the EURO), it is verbatim showing what the model outputted. Different types of maps may result in slightly different outputs, as they use different formulas to decide how much snow has fallen (or stuck, in some cases)
    1. sometimes i or others post percentile maps - those are also only for ensembles. These are like the confidence maps
  2. I don't think it's that easy but I'm the wrong guy. I think they try and knock out these biases in subsequent updates but the best forecasters almost like to know the biases so they know how to correct or mentally adjust. YMMV - idk which biases are real these days
  3. I think most of us would consider EURO king but GFS has it's days (yesterday). NAM did well yesterday too... some storms seem to work out better for the American suite of models then others. the GEM/CMC and UKIE are supposedly good models but maybe not for our specific purposes.
Link to comment
Share on other sites

I've mostly lurked here over the past several years,  just posting my observations during snow events.  I have a couple of general questions about models, don't know if this is the right thread but don't want to post on a subforum that's dead with no chance of a response.
(1) When models spit out the contoured snow maps, do the figures within those maps represent a mean projected accumulation (i.e. if its 6 inches, that means a deviation of +/-  2" depending on confidence level), or do those totals represent the maximum projected accumulation (i.e. not more than 6", within confidence level of 95%). I know these maps are always a snapshot in time, and some maps are specific about what they're showing, such as the 10% chance/confidence maps.  But if not, are the figures just the mean?
(2) I often hearing that ______ model has a north bias, a south bias, generally projects higher qpf totals, etc.  If this is in fact true, meaning the bias is statistically significant, why aren't those models corrected to account for the bias? After all, aren't all of these models based on data or other markers from past storms, compared with incoming data from the current storm? And the purpose of any statistical model is to become progressively more accurate over time? Or are the people claiming real bias just weenies?
(3) Everyone seems to have slightly different opinions, but as far as model supremacy is concerned, how do each rank at various stages of a storm (168 hours, 98 hours, 48 hours, 24 hours, 12 hours).  Based on what I've observed, there seem to be three models that lead the pack, tell me if I'm right:
  • EURO-- Seems to be the consensus best model at all stages
  • GFS-- Seems to be about the same as the NAM, but
  • NAM-- Seems to given more attention at about the 24 hour stage
  • Then....  ICON, GEFS, UKIE, HRRR, GDEPS, GEM, CMA, etc.--  all works in progress.
Any input would be appreciated..  Thanks!!

I’m not the one who should answer, but the questions are interesting and I’m waiting on sushi rn. For #2 I think that’s where ai/ml would add value (since it should be based more so on what usually occurs instead of what should) and definitely agree on #3 regarding the euro. It just seems like an all around dominant model.
Link to comment
Share on other sites

1 minute ago, midatlanticweather said:

AI is so resource-intensive! Need some more hamsters on wheels in the basement. 

Actually the Euro AI just achieved consciousness and launched a missile strike against the computers running the regular Euro.

  • Like 1
  • Haha 2
  • 100% 1
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...