Jump to content
  • Member Statistics

    17,584
    Total Members
    7,904
    Most Online
    LopezElliana
    Newest Member
    LopezElliana
    Joined

Big Snow threat, what will it do, part II


Typhoon Tip

Recommended Posts

  • Replies 1.6k
  • Created
  • Last Reply

Say what?  First of all, little known fact that the GFS analysis resolution is actually HIGHER than the ECMWF analysis resolution (the analysis increment is done on a grid roughly 2x the ECMWF grid I think).  This is because 4DVAR is so much more expensive than the Hybrid 3D EnVar system that is used in the GFS/GDAS. 

 

What exactly did you read and what are you referring to?  The GDAS system also has many noise/smoothing constraints within it, including an incremental normal mode initialization technique.  I'm happy to pass you a reference if you're interested.  Now, given that the NCEP system is still 3D, there are issues, such as treating the observations at their correct time, extracting tendency information from the observations etc.  NCEP is working aggressively to implement a 4D variant for the GDAS....not the uber expensive 4DVar that requires the tangent-linear and adjoint models, iteratively as part of its solver, but instead an extension to the current operational hybrid called 4D EnVar.  Preliminary results are extremely encouraging.

 

Lastly, in terms of the impact experiments you cite, they aren't particularly meaningful.  Data assimilation is performed incrementally, so by attempting to drop the ECMWF analysis into the GFS model, you are actually injecting features from the ECMWF model itself.  NCEP has yet to glean anything meaningful and practically relevant about the operational system from such injection experiments.

 

I'm happy to have a discussion on the merits of 4DVAR, EnKF, and hybrid variants, as well as a provide my opinion on the cost-benefit analysis if you are interested in learning more.

 

I was spending time here, http://www.ecmwf.int/research/ifsdocs/CY28r1/Assimilation/index.html, and they provide a fair enough amount of information about the history/evolution of the 3-D variable system, and the subsequent migration to the 4-D system for assimilations at ECMWF, there differences/usefulness.  

 

As far as the grid spacing .. the same site lists the resolution table:



Spectral Trunc.	Gaussian Reduced	Lat/Long	Application
T1279	N640	0.125	Deterministic forecast + data assimilation
T639	N320	0.25	EPS up to 10 days
T319	N160	0.5	EPS from 10 to 15 days
T319	N160	0.5	EPS from 16 to 32 days (in monthly forecast only)
T255	N128	0.75	Seasonal forecast system 4
-	-	0.25	Deterministic ocean wave
-	-	0.5	EPS ocean wave
T255	N128	0.75	ERA Interim
T159	N80	1.125	ERA 40

NCEP site list the "master" GFS as at .5 degrees resolution, as provided here:  http://www.nco.ncep.noaa.gov/pmb/products/gfs/

 

But,

T1279 N640 0.125 Deterministic forecast + data assimilation

 

Reads like the Euro has a finer mesh.. when comparing to the .5 GFS, above.  I am not sure how to atone for any differences therein, or even if that interpretation is correct.   

Link to comment
Share on other sites

this is such a tantalizing look

f48.gif

can anyone please explain to me why there is that donut hole on the h7 RH maps (translating down to the surface)???

 

Looking at the h5, I would think there would be a low getting ready to deepen close to the coast

Link to comment
Share on other sites

can anyone please explain to me why there is that donut hole on the h7 RH maps (translating down to the surface)???

 

Looking at the h5, I would think there would be a low getting ready to deepen close to the coast

2 different things going on. You have the initial overrunning from I90 through New England...then the developing secondary off the SE coast.

Link to comment
Share on other sites

this is such a tantalizing look

f48.gif

 

I'm not sure it will ultimately verify like that or not but if we had 100 setups like that, maybe 3 turn out as bad as this one is currently modeled, there is always the unusual case but odds are against it being as bad as its been shown so far.

Link to comment
Share on other sites

I'm not sure it will ultimately verify like that or not but if we had 100 setups like that, maybe 3 turn out as bad as this one is currently modeled, there is always the unusual case but odds are against it being as bad as its been shown so far.

and aside from the last 13 years, this is why KU's are special events.....lots of stuff needs to go right

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...