Jump to content
  • Member Statistics

    17,589
    Total Members
    7,904
    Most Online
    LopezElliana
    Newest Member
    LopezElliana
    Joined

Jan 23/24 Major Coastal Storm Discussion


Zelocita Weather

Recommended Posts

  • Replies 2.9k
  • Created
  • Last Reply

Possible yes... Likely no, if this was a 1 or even 2 model storm to this point, I'd say yes... Model concensus is overwhelming, the pattern is ripe, all the ingredients are there in terms of High pressure weakening and a closed of low... No reason at this point IMO to think this will go OTS.. Not saying it isn't possible, but there isn't any evidence to support that solution

 

 

Model consensus certainly would seem to make a major shift in storm evolution less likely, although it's always possible that some of the key initial conditions, which are critical to model accuracy (and whose errors propagate through every forward time increment in the models), are either inaccurate or were poorly sampled - and since these ICs are presumably all identically fed into every model, one could imagine all the models see the same resulting error.  Unlikely, but possible.  

 

I posted this yesterday, but received no comments.  It's at least interesting that last night's models and this morning's models, presumably all fed identical (or close to it) initial conditions, provided near consensus last night vs. widely divergent outcomes today.  Obviously, the physics, chemistry and thermodynamics (and key assumptions about each) used in the models are somewhat different, as are the numerical methods used to "solve" the large number of resulting differential equations over time, since these non-linear systems can't be truly "solved" mathematically and require numerical approximations (as are grid spacing and numerical interpolation routines to flesh out the 3-D space) - the question is why such a significant divergence in outcomes today vs. yesterday?  

 

Any pros out there care to comment?  Also, at 0Z tonight, will we know which model predicted the 12 hours from 7 am to today to 7 pm tonight best and would we then say that model X was garbage, but model Y was spot on?  Do they do diagnostics on how each model run performed, for the next 12/24/etc. hours?  Pretty sure I've asked this before, but not sure I got an answer...

Link to comment
Share on other sites

I posted this yesterday, but received no comments. It's at least interesting that last night's models and this morning's models, presumably all fed identical (or close to it) initial conditions, provided near consensus last night vs. widely divergent outcomes today. Obviously, the physics, chemistry and thermodynamics (and key assumptions about each) used in the models are somewhat different, as are the numerical methods used to "solve" the large number of resulting differential equations over time, since these non-linear systems can't be truly "solved" mathematically and require numerical approximations (as are grid spacing and numerical interpolation routines to flesh out the 3-D space) - the question is why such a significant divergence in outcomes today vs. yesterday?

Any pros out there care to comment? Also, at 0Z tonight, will we know which model predicted the 12 hours from 7 am to today to 7 pm tonight best and would we then say that model X was garbage, but model Y was spot on? Do they do diagnostics on how each model run performed, for the next 12/24/etc. hours? Pretty sure I've asked this before, but not sure I got an answer...

I disagree with 1 main point... only 1 model was "widely divergent" and the euro wasn't even all that much different from it 00z, continued to show strong confluence as it has.. The GFS was off slightly on the 06z but came right back full force with 12z.... And the GGEM, as I've been saying for days, has not moved since last Friday.. Same output.. 2 times a day lol

Link to comment
Share on other sites

One thing to emphasize during the PBP.  Despite just about everyone's assumptions, there's NO evidence that trend is relevant in assessing model output.  I.e., the best way to make a forecast is to consider all the model runs you have, weighting more recent runs accordingly, but you don't see any signal by giving any weight to "trend". 

 

I.e., take 'trend' for what it really is - illusory signal that just happens to be random walk lined up in a row.

Link to comment
Share on other sites

Archived

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

Guest
This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.

×
×
  • Create New...